Message Boards Message Boards

10
|
22985 Views
|
302 Replies
|
227 Total Likes
View groups...
Share
Share this post:
GROUPS:

[WSG24] Daily Study Group: Programming Proficiency

A Wolfram U Daily Study Group previewing our upcoming Programming Proficiency course sequence begins on Monday, January 29, 2024. The study group will run for ten days through February 9, and each day will run from 11AM to noon CDT.

Join me, @Abrita Chakravarty and a cohort of fellow enthusiasts to build a strong foundation for any and all sorts of Wolfram Language programming that you'll do in the future. We'll cover everything from syntax, definitions, and the underlying structure of expressions to efficient coding, working with various data structures, and even developing packages.

This study group will cover material from our Programming Fundamentals and Practical Programming courses, as well as provide attendees a first look at Programming and Development, the upcoming third and final course in the Programming Proficiency course sequence. Our intent is to provide a setting for this material which not only reaps the benefits of our Daily Study Groups'... well, group nature, but also spreads this material out over two weeks for those who cannot commit to the three-hour blocks required by these courses in their natural habitat.

This study groups aims to get you up and running with Wolfram Language, and as such no prior Wolfram Language or programming experience is necessary.

REGISTER HERE.

We look forward to seeing you there!

enter image description here

POSTED BY: Arben Kalziqi
302 Replies

Hi folks! We've been enjoying your Level 2 certification submissions so far. If you're curious about whether an idea you have would be fit to submit, please feel free to ask in this thread. (If you have technical questions, however, please wait for the next Programming Proficiency Office Hour session, which is currently scheduled for July 25.)

POSTED BY: Arben Kalziqi

Hi Arben,

Thank you for the quiz update.

Problem 7 is broken: none of the choices yields a correct answer. I know which answer I like best though, teehee!

Lori

POSTED BY: Lori Johnson

Thanks for pointing this out, Lori! When I updated the question, the number of answer choices changed and messed up the key—but I'd failed to notice it. It should be fixed soon.

POSTED BY: Arben Kalziqi

Thank you, Arben!

I hope you like my function. I'm still adding to it!

POSTED BY: Lori Johnson

Could you please advise if this program would be suitable for me if I only know how to set up and configure development environments and have just started learning programming languages?

POSTED BY: Ellis Miller
Posted 10 months ago

This course may not be the place to start. If I were starting today, I'd take the Elementary Introduction to the Wolfram Language Interactive Course. This is a big course that covers a lot of different areas of programming and the Wolfram Language. Work all the exercises. It may take several months to get through it, but you'll be left with a good fundamental knowledge of the language after completing. The course text is also available as a physical book, but I don't think the book is the best way to learn this material. Even if you get the book, going through the examples and quizzes online will be required. If your learning style is different and you do get the book, make sure you get the Third Edition. Also, the content of the book is available for free online (follow a link on the book link above).

Get in the habit of looking through the Wolfram Language Reference available online. Notice that you can have a bunch of windows open simultaneously into Wolfram Language code and documentation. You can even do "open frame in new tab" if you need more real estate in one of the sub-windows of the EIWL course. Most importantly, have fun! Write original code about things you know and things you're interested in.

POSTED BY: Phil Earnhardt

Understood, thank you for the advice. Special thanks for "Most importantly, have fun! Write original code about things you know and things you're interested in."

POSTED BY: Ellis Miller
Posted 10 months ago

Replacement rules.

{4, -2, 3.14, 9, -25, 16} /. x_ /; x > 0 && IntegerQ[x] :> Sqrt[x] from the lesson I get.

But I get stuck with the following :

listA= {{a, b, a}, {b, a, a}, {a, b, a, b}, {a, b, b, a, a}, {b, a, b, a}, {b, b, a}, {a, b, a, b}, {b, a, b, a, b,},{ a, a, a, a}}

listA /. {a -> b, b -> a}. This is no problem

But now I want the replacement ONLY to apply to the sublists that start with a ‘b’.

I tried with quite some different ways, but all seem to fail :

listA /. {a -> b, b -> a} /; _First == b.
I get : ReplaceAll::reps: {{a->b,b->a}/;_First==b} is neither a list of replacement rules nor a valid dispatch table, and so cannot be used for replacing.

I tried with a pure function and got nowhere.

Finally, this inelegant creature at least works.

Table[If[SameQ[listA[[n, 1]], b], listA[[n]] /. {a -> b, b -> a}, listA[[n]]], {n, 1, Length[listA]}]

How could I do I do it more ‘Wolfram’ style? If possible, show me how to do it with a pure function as well.

Thanks for your suggestions.

POSTED BY: B. Cornas
Posted 10 months ago

I would do something like this

Cases[listA, x_List /; First[x] == b]

{{b, a, a}, {b, a, b, a}, {b, b, a}, {b, a, b, a, b}}

Cases[listA, x_List /; First[x] == b] /. {a -> b, b -> a}

{{a, b, b}, {a, b, a, b}, {a, a, b}, {a, b, a, b, a}}

POSTED BY: Doug Beveridge
Posted 10 months ago

Thanks for your suggestion, Doug.

I might not have been too clear in my question. But actually I need to get the whole listA back, with only the sublists starting with 'b' changed. But in the sublists that are changed, all element should change.

So sublist {b, a a,b} should become {a,b,b,a}. My apology if I was not clear enough. And for the complete ListA =

{{a, b, a}, {b, a, a}, {a, b, a, b}, {a, b, b, a, a}, {b, a, b, a}, {b, b, a}, {a, b, a, b}, {b, a, b, a, b},{ a, a, a, a}} ,

the should get the result :

{{a, b, a}, {a, b, b}, {a, b, a, b}, {a, b, b, a, a}, {a,b,a,b}, {a,a,b}, {a, b, a, b}, {a,b,a,b,a},{ a, a, a, a}}

POSTED BY: B. Cornas
Posted 10 months ago

blank

POSTED BY: Doug Beveridge
Posted 10 months ago

OK

listA /. x_List /; First[x] == b :> ReplaceAll[x , {a -> b, b -> a}]

{{a, b, a}, {a, b, b}, {a, b, a, b}, {a, b, b, a, a}, {a, b, a, b}, {a, a, b}, {a, b, a, b}, {a, b, a, b, a}, {a, a, a, a}}

POSTED BY: Doug Beveridge
Posted 10 months ago

Thanks Doug. Works great.

I wonderful (and sometimes confusing) to see in how many different ways something can be coded in Wolfram.

POSTED BY: B. Cornas
Posted 10 months ago

Thats why I loved this course by Arben . I suddenly saw things from many angles and it started to make more sense .

eg

f[p_ /; (IntegerQ[p] && p > 0)] := p f[p - 1]

f[p_] := p f[p - 1] /; Positive[p] && IntegerQ[p]

f[p_Integer?Positive] := p f[p - 1]

f[p_?(IntegerQ[#] && Positive[#] &)] := p f[p - 1]
POSTED BY: Doug Beveridge
Posted 10 months ago

Yes but it is still confusing

listA /. x_List /; First[x] == b :> ReplaceAll[x , {a -> b, b -> a}]

{{a, b, a}, {a, b, b}, {a, b, a, b}, {a, b, b, a, a}, {a, b, a, 
  b}, {a, a, b}, {a, b, a, b}, {a, b, a, b, a}, {a, a, a, a}}

works
but

listA /. x_List /; First[x] == b :> x /. {a -> b, b -> a}

{{b, a, b}, {a, b, b}, {b, a, b, a}, {b, a, a, b, b}, {a, b, a, 
  b}, {a, a, b}, {b, a, b, a}, {a, b, a, b, a}, {b, b, b, b}}

does not .......

Maybe Arben can tell us why ?????

POSTED BY: Doug Beveridge

I think it's just order of operations/priority stuff—if you use parentheses like you see in my other post, it works just fine. I didn't actually test it, just wrote it how I figured it was most likely to work!

POSTED BY: Arben Kalziqi
Posted 10 months ago

Thanks Arbin

Yes it is the Parenthesis (good to know )

listA /. x_List /; First[x] == b :> (x /. {a -> b, b -> a})

{{a, b, a}, {a, b, b}, {a, b, a, b}, {a, b, b, a, a}, {a, b, a, 
  b}, {a, a, b}, {a, b, a, b}, {a, b, a, b, a}, {a, a, a, a}}
POSTED BY: Doug Beveridge
Posted 10 months ago

Ditto. Arben gives new and very insightful perspectives on the material. He teaches on many levels simultaneous and his examples are really exemplary. A very complete teacher, which is quite rare.

POSTED BY: B. Cornas
Posted 10 months ago

The issue is here:

/; _First

The /; means that you write some test afterward—not only is _First not a component of a test, but it's not the correct pattern syntax. Recall that what follows a Blank (_) should be a head to be matched; i.e. you can't put an arbitrary function there and take it to be acting on whatever the blank is.

If you have

listA= {{a, b, a}, {b, a, a}, {a, b, a, b}, {a, b, b, a, a}, {b, a, b, a}, {b, b, a}, {a, b, a, b}, {b, a, b, a, b},{ a, a, a, a}} (*i removed the errant comma*)

and want to do {a -> b, b -> a} on any sublist that starts with b, you could do something like

listA /. l_List /; First[l] == b :> (l /. {a -> b, b -> a})

if you wanted to use Condition. Better yet, just match the pattern directly:

listA /. {b, rest__} :> ({b, rest} /. {a -> b, b -> a})
POSTED BY: Updating Name

(this is arben.)

POSTED BY: Arben Kalziqi
Posted 10 months ago

Hi Updating Name (Rohit?),

I suspected that I did not use First correctly :-)

Your solution works excellent. Also if I nest the sublists deeper and irregular (different nesting levels). That is great.

Do I understand it correctly, if I say :

{b, rest__} looks at all the innermost sublists and if there is such a sublist that starts with a 'b' and has one or more elements, then the replacement takes place?

Thanks again for your valuable help,

Best, Bert

POSTED BY: B. Cornas

yes, exactly—if you have a list that starts with b and has one more or elements after the b, then what I wrote will work. If you want to include {b} alone, you have the exact right idea with ___/BlankNullSequence.

POSTED BY: Arben Kalziqi
Posted 10 months ago

You reply faster than I can edit. Wow :-)

POSTED BY: B. Cornas

Arben, New general question -- do you know what is going on with Demonstrations? In anticipation of getting my Wolfram Level 2 certificate, I have submitted a demonstration, but have heard nothing back. The formal submittal process seems to be broken, and attaching the Demonstration notebook to an email (to the Wolfram Demonstrations Group) didn't seem to elicit a response either. I have submitted a question on the matter to Wolfram Tech Support, but have heard nothing back (yet). Can anyone tell me if the Demonstrations project is still active? Finally, will time run out on my Level 2 certificate before the issues are resolved? Sorry to bother you with this, but I would like to know what is going on just to ease my mind that I have done all I can. Thanks, Mike

POSTED BY: Michael Ulrey

Hey Mike, we'll get back to you soon about this. In the meantime, please don't worry about any timeline issues—as far as we're concerned, you've submitted your project within the deadline and that's that.

POSTED BY: Arben Kalziqi

Arben, I still have not heard anything about the Level 2 certificate. Maybe it just takes awhile, and I can understand that. However, I'm afraid I may have gummed up the works when trying to pay the fee for Level 2 using the web payment software. I got an "unexpected error" message, and the software then interpreted the discount code as my payment (several times). Christine helped me get my money back, but this left me wondering if the system thinks I have not paid at this point. So I just want to make sure that I'm still in the queue to get a certificate, or if I need to do something more. Thanks.

P.S. In the meantime, an attendance certificate for the AI webinar (which I took before the Programming Proficiency course) showed up on my LinkedIn account. I'm worried that the system thinks this is what I was trying to get credit for, and has now dropped the Level 2 request (associated with the Programming Proficiency course and subsequent Demonstration submittal).

POSTED BY: Michael Ulrey

Hey Michael—I'm sorry about that! I finished grading it 2-3 weeks ago; I'm not sure what got blocked up in the process but we're on it. (Spoiler: you passed)

Edit: you should have it later today.

POSTED BY: Arben Kalziqi

Thanks!

POSTED BY: Michael Ulrey
Posted 10 months ago

Right, thanks Arben :-)

I just got my first connection through a chat Nb with chatGPT and asked the same question. After two misses, and restating my question more precise (and partly double), I gotv a working answer :

Map[If[First[#] === b, # /. {a -> b, b -> a}, #] &, listA]

It's nice to see it work, although your solution is more refined, I think, and probably faster :-)

But the A.I. solution shows me again the pure function, which I understand in theory, but in practice I often seem to do it wrong. It will come, some day.

Also, in the If statement (which I also used, I first used == instead of what I should have done : ===. I dodged my error by using SameQ[]. So I think this A.I. coding help can be useful in the learning process.

POSTED BY: B. Cornas

Yeah, it can definitely be useful for exploring—it can also definitely be a little frustrating for reasons you touch upon, but so it goes :)

POSTED BY: Arben Kalziqi
Posted 10 months ago

Arben, I tried using Timing[...] and AbsoluteTiming [...] on 3 different instructions to accomplish the same thing: Power, an anonymous function using Keys, and an anonymous function using indexing. I used a large enough list of associations so that Timing was non-zero. All three had Timing between about 2.1 and 2.8 seconds, and AbsoluteTiming between about 3.7 and 3.9 seconds. I ran this in Mathematica 14 on an HP Envy running Windows 11 Home 64-bit OS. My notebook is attached as file TimingExample_1mar24.nb .

I expected Power to be clearly faster but it wasn't. Please explain these timing results.

POSTED BY: Gerald Dorfman
Posted 10 months ago

give me your timing for a much smaller problem .

{<|"a" -> 1, "b" -> 2|>, <|"a" -> 3, "b" -> 4|>, <|"a" -> 5, "b" -> 6|>}

POSTED BY: Doug Beveridge
Posted 10 months ago

Using

shortListOfAssoc = {<|"a" -> 1, "b" -> 2|>, <|"a" -> 3, "b" -> 4|>, <|    "a" -> 5, "b" -> 6|>}; 

I added Timing, AbsoluteTiming, and Repeated timing to the file I uploaded before, TimingExample_1mar24,nb , which is attached. (I also added RepeatedTiming for listOfAssoc that I ran previously.) My respective results for shortListOfAssoc using Power, anonymous function with tags, and anonymous function with indices are as follows:

Timing: all three were 0.

AbsoluteTiming: 0.0000124, 0.0000111, 0.0000101

RepeatedTiming: 1.3700510^-6, 2.1609110^-6, 4.3709*10^-6

POSTED BY: Gerald Dorfman

Looking into this and will let you know what I find. Something interesting is that the timing is pretty nonlinear even for an operation like this, so I expect there's something going on with how the values are stored under the hoot.

POSTED BY: Arben Kalziqi
Posted 10 months ago

Hi Arben,

I could not properly formulate my question in The Office Hours Session as I had only joined in at the very end (a miscalculation on my part). So thank you for suggesting I pose my question here.

I am trying to make a kind o’f Rhythm Evolver’ by means of an Evolutionary process. One thing is that I need to detect the degree of ‘regularity’ in a rhythm as a fitness score parameter.

This way I also try to find out if Nesting as a ‘regularity’ in music is easily, or at all recognised by listeners. Nesting is the most complicated and hidden regularity. Repetitions are much easier recognised.

Let’s just talk about just one line (one rhythm instrument, e.g. Kick drum), as I think it is already difficult enough.

A hit is 1, no hit is 0.

  1. I have already been given a way to detect ‘Repetition’, ‘Symmetry’ and ‘Permutations’ of parts (also here in the Community :-)

Now I am looking for a way to detect ‘Nesting’.

E.g. The following 3 parts : {1,0,0,1} {1,0,0,1,0,1,1,0} {1,0,0,1,0,1,1,0,0,1,1,0,1,0,0,1}
all come from the nesting Rule : 1->{1,0}. and 0-> {0,1} starting with {1,0} and applied respectively 1x, 2x and 3x.

I want to go the other way, and see if in a rhythm line, that has been evolved , there is any sort of Nesting as regularity. Not with these specific rules and specific initial condition. I want to detect ANY Nesting (any rules, any initial condition).

So, given the line { 1, 0, 0, 1, 1, 0, 0, 0, 1, 1, 0, 1, 1, 0, 0, 1}, I should detect that the bold part (1, 0, 0, 0, 1, 1, 0, 1, 1) is a nested part from:

               Initial conditions {1,0} with the rules : {1 -> {0, 1, 1}, 0 -> {1, 0, 0}}

It is no really important that I detect what rules and which initial condition, but I want to detect which part of my Rhythm Line is a Nested part in any form.

I am not sure if it is possible, without trying all the rules. This due Stephen Wolfram’s ‘Computational Irreducibility’ .

What are your thoughts?

Thanks, Bert

POSTED BY: B. Cornas

Hi Bert—hmm... my first thought is that in general your feeling about computational irreducibility is correct. However, if the cases are 1:1 like you've shown so far (which I think you say that they aren't), then you can certainly do it a few ways.

  • Less elegant but probably most efficient is to partition your sequence up by rule size—if the nesting creates three numbers, you can just partition a result into blocks of 3 and reverse the nesting rules to get back to where you started.
  • More elegant but less efficient (because it requires pattern-matching that could be rather general) would be to use SequenceReplace.

But again, if the rules aren't 1:1 you're going to run into issues. If they output different lengths, you'll need to use SequenceCases and you may be able to work backwards from there.

POSTED BY: Arben Kalziqi
Posted 10 months ago

Thanks Arben. Do you mean by "if the cases are 1:1 . . ." that the rules are of equal length?

So {1->{1,0,1} and 0->{0,0,1}} is 1:1 and {1->{1,0,1} and 0->{0,1}} is Not?

Reverse the nesting rules is not quite possible, I guess, as I do not know which rules might have been used. Also, the nesting might be 2 or 3 deep, so I do not know the length to parse. It feels a bit like backwards running a C.A., which is possible in some cases, but not very often.

Checking for repetitions in a rhythm line is easy, but this Nesting looks like a different animal :-)

As the computation comes into a repeated process of measuring the amount of Nesting in a rhythm ( so thousands of these detection measurements), it might be an idea to create a lookUp table with all the possible nestings for two elements (0 and 1). If a take as a max of number of initial elements, as well max number of elements in a rule to be 3 (which I think servers my purpose) then there would be 14 possible initial conditions and 12 rules for 0 and twelve rules for 1. As these rules are 'mirrored, I need only to take into account half of them. Then I create the lookup table for all Nestings from level 1 to say 3, which should be enough. From there on I could use SequenceCases or SequencePositions.

This will still leave the challenge for detecting Nesting in a complete rhythm, where I use 16 possible values for the possible simultaneous occurances of hits from 4 instruments together. This will explode with lookUp tables, I am afraid.

Best, Bert

POSTED BY: B. Cornas

By 1:1 I mean it in the sense of functions—as in, "in some given sequence, can any one particular output subsequence only be generated by one input?" If you don't have that, I don't think it's feasible.

POSTED BY: Arben Kalziqi

What is the deadline for using the Level 2 discount code I received? (I took the first WSG24 this year)

I have a project idea but haven't had much time since the course ended and probably won't for a couple weeks longer.

POSTED BY: David Snyder

@David Snyder, the $50 promo offer is good through June 30, 2024.

POSTED BY: Jamie Peterson

Never mind, (content removed)

POSTED BY: Carl Hahn
Posted 10 months ago

I am looking for a way to do an operation inside a nested List. The nesting is 4 deep and the inner-lists are of unequal length.

E.g. listA = {{{{1, 2}, {3, 4}}, {{5, 8}, {9, 12}, {14, 17}}}, {{{11, 13}, {15, 17}}, {{18, 20}, {22, 24}}, {{26, 31}}}}

I want to subtract the first number form second number in each sub-list. The nesting needs to be preserved.

So the result should be : {{{{1}, {1}}, {{3}, {3}, {3}}}, {{{2}, {2}}, {{2}, {2}},{{5}}}}

I tried with Map at level{3} and {4} but it did not work. Or with a pure function, also no dice. Also Table. I tried quite a lot of things, but the deep nesting is throwing a wrench into my engine.

POSTED BY: B. Cornas

Today's lecture should do you well :).

In[53]:= 
listA = {{{{1, 2}, {3, 4}}, {{5, 8}, {9, 12}, {14, 17}}}, {{{11, 13}, {15, 17}}, {{18, 20}, {22, 24}}, {{26, 31}}}};
listA /. {x_Integer, y_Integer} :> {y - x}

Out[54]= {{{{1}, {1}}, {{3}, {3}, {3}}}, {{{2}, {2}}, {{2}, {2}}, {{5}}}}

More to the style you'd imagined—and faster for a large list—is:

In[57]:= Map[{-Subtract @@ #} &, listA, {-2}]

Out[57]= {{{{1}, {1}}, {{3}, {3}, {3}}}, {{{2}, {2}}, {{2}, {2}}, {{5}}}}

And the best version, I think:

``` In[69]:= Apply[{#2 - #1} &, listA, {-2}]

Out[69]= {{{{1}, {1}}, {{3}, {3}, {3}}}, {{{2}, {2}}, {{2}, {2}}, {{5}}}}```

POSTED BY: Arben Kalziqi
Posted 10 months ago

Wow, even after what must be an exhausting session for you as well, I get an immediate answer. Truly amazing :-)

  1. I tried also with Rules, but used a too wide a pattern (I cannot put it here as the 'underscore' disappears in the post. I get my mistake.
  2. The map example I have to study tomorrow. I am also tired now
  3. In Apply[{#2 - #1} &, listA, {-2}]. I missed the function at first, but of course it is the Subtract (the -). So I rewrote the function :

    Apply[Subtract[#2, #1] &, listA, {-2}]
    

and strangely enough the innermost level is gone, which by the way is excellent for me in this case. So I get now {{{1, 1}, {3, 3, 3}}, {{2, 2}, {2, 2}, {5}}}

Why is there a difference with Apply[{#2 - #1} &, listA, {-2}]?
I looked at FullForm and Trace, but could not deduct why it is different.

Thanks a million, Arben,

Best, Bert

POSTED BY: B. Cornas

Hi Bert—I thought you wanted the extra braces still for some reason, so I had added them back in! They disappear with Apply because Subtract (and indeed -) wants two arguments provided as a sequence, not as a list—that's why we use Apply, because Apply[Subtract,{a,b}]===Subtract[a,b]. Hence, you can see the difference between Apply[Subtract[#2,#1]&,...] and Apply[{Subtract[#2,#1]}&,...].

I see why writing the Subtract out is helpful for clarity, though I would say that getting an intuitive sense of what something like #2-#1&@@ is really doing is indispensable to working efficiently with Apply and its siblings. I would say you can think of it as "an operation which strips out the second and first arguments of ANY structure and gives me their difference". It works on lists, it works on sums and products, it works on any arbitrary g[a,b] with the action that #2-#1&@@g[a,b]===b-a.

Also, you can type code between backticks here, like `hello_` in order to prevent any unwanted markdown formatting. You can start a new line with 3 backticks, press enter, then type/paste code and do another new line with 3 backticks in order to get the formatted boxes you see in my previous reply.

POSTED BY: Arben Kalziqi
Posted 10 months ago

Great Arben, I get it.

So, to get rid of the unwanted brackets, I could do :

Apply[#2 - #1 &, listA, {-2}] or Apply[#2 - #1 &, listA, {3}]

Both give the same result. Is there a hidden advantage in using levels counting from the back {-2} ?

To get my end result in this, I have :

Total[Apply[#2 - #1 &, listA, {3}], {3}]
This sums the partial results and that is what I need to continue.

Now the Map from yesterday.

Map[Subtract @@ # &, listA, {-2}]

I think I understand it, but find it hard to come up with myself. I tried to rewrite it without the 'shorthand', just to get a grip on it.

`Map[Apply[Subtract[#[[2]], #[[1]]] &, listA], {-2}]

Map[Apply[Subtract[#[[2]], #[[1]]] &, listA, {-2}]]`

Both are not correct. It looks as if the level {-2} is taken only after the Apply is done. Still , the second example, where I moved the level-spec into the Apply, it still does not work.

I am getting closer to understanding it, but I am not quite there, I am afraid.

Arben, your help is really great :-)

Best, Bert

POSTED BY: B. Cornas

Since you're Apply-ing, you get rid of the List structure. In that context, what does #[[2]] or #[[1]] mean?

POSTED BY: Arben Kalziqi
Posted 11 months ago

...

POSTED BY: lara wag
Posted 11 months ago

Suppose there is an external database or other data source. Unfortunately, in this external data source, the keys are in lower case with underscores ("snake_case"), e.g. company_name. To work properly in Wolfram Language, the keys must be in camel case, e.g. companyName.

Here is an example that works on a List of Associations:

AssociationThread[
   StringReplace[Keys[#], 
     MapApply[Rule, 
      Transpose@{"_" <> # & /@ Alphabet[], 
        ToUpperCase[Alphabet[]]}]] -> Values[#]] & /@ data

To be honest, from my perspective it looks ugly, especially the "Transpose" part. Any ideas on how to make this transformation more concise?

... or is there a built-in function for this that I don't know about? I saw this resource function https://resources.wolframcloud.com/FunctionRepository/resources/ToCamelCase/, but I wasn't able to make it run in this "key of association" problem.

Notebook with sample data attached.

POSTED BY: lara wag

Hi Lara—keys in an association can be basically whatever you want, and certainly be in the form of "company_name" for example. If you need to access them via "slot" notation, you can just wrap them in quotes:

In[280]:= #"company_name" &@<|"company_name" -> 2|>

Out[280]= 2
POSTED BY: Arben Kalziqi
Posted 11 months ago

Thank you, Arben, for your super-fast reply.

Interesting - so it works with quotation marks. - Thank you!

POSTED BY: lara wag
Posted 11 months ago

Hi Lara,

An easier way to rename the keys is to use KeyMap.

data // Map[
  KeyMap[StringReplace[{"_id" -> "ID", 
                        "id" -> "ID",
                        "_" ~~ c : LetterCharacter :> ToUpperCase@c}]]]
POSTED BY: Rohit Namjoshi
Posted 11 months ago

Thank you very much, R, impressive! I was looking for something like that.

POSTED BY: lara wag

Hi Arben;

I remember you covering in one of the lectures a technique of applying MapAll (/.) before evaluating the expression as opposed to after the expression is evaluated. However, I cannot remember the way you formatted the function. Can you please show me one more time? The example below shows the values being assigned after the function is evaluated, which is not what I want. What I want are the values of Sin[t], Cos[t], Tan[t] to be assigned to x, y, z before any other functions are evaluated.

Thanks,

Mitch Sandlin

D[Log[x^2 + y^2 + z^2], t] /. {x -> Sin[t], y -> Cos[t], z -> Tan[t]}
POSTED BY: Mitchell Sandlin

I think you just do exactly what you said you want to do, i.e. assign the variables before you take the derivative:

D[
Log[x^2 + y^2 + z^2] /. {x -> Sin[t], y -> Cos[t], z -> Tan[t]}
, t]

Or even before you take the Log, but I think that gives you the same answer:

D[Log[(x^2 + y^2 + z^2) /. {x -> Sin[t], y -> Cos[t], z -> Tan[t]}],
  t]
POSTED BY: Carl Hahn
Posted 11 months ago

Hi Mitchell

/. is ReplaceAll, not MapAll. I don't recall what Arben showed but here are a couple of ways

Unevaluated[D[Log[x^2 + y^2 + z^2], t]] /. {x -> Sin[t], y -> Cos[t], z -> Tan[t]}

Hold[D[Log[x^2 + y^2 + z^2], t]] /. {x -> Sin[t], y -> Cos[t], z -> Tan[t]} // ReleaseHold

(* (2 Sec[t]^2 Tan[t])/(Cos[t]^2 + Sin[t]^2 + Tan[t]^2) *)
POSTED BY: Rohit Namjoshi

Take Rohit's answer, but also note that if you were to write x'[t] and so on—that is, explicitly note the time dependence—D will take those derivatives symbolically and you can plug in values for the time-dependent functions after the fact, at which point the derivatives will be evaluated.

POSTED BY: Arben Kalziqi

Arben, I don't think I understand what you just said (a rare thing...). How would you re-write the expression?

I liked the elegance of the "Hold" and "Release Hold" solution, but only because it's explicit. However, I do get the same answer by using ReplaceAll within the brackets of the Derivative forcing it to happen before the derivative is taken. I've noticed in other people's examples that you can stick the /. operation all over the place. What is wrong with that?

(my own answer would be that if you are writing longer code, it gets hard to read and understand - I'm just asking a more generic question about WL grammar rules) Carl

POSTED BY: Carl Hahn
Posted 11 months ago

I've noticed in other people's examples that you can stick the /. operation all over the place. What is wrong with that? (my own answer would be that if you are writing longer code, it gets hard to read and understand - I'm just asking a more generic question about WL grammar rules) Carl

My suggestion is to use the standard functional call (ReplaceAll[], etc.) until you thoroughly understand them. You should always -- always -- be able to see and say the long form when the shortcut is being used. IMHO, it's a big problem to use shortcuts if you don't always have the formal name of the function call on the tip of your tongue. Note: even things like _x have a standard form: Blank[x]. Computers have the ability to expand any shortcut within a microsecond; we should be ability to translate either way with a similar amount of minimal computational overhead.

You want to be able to rattle off a spell with the ease of Hermione Granger and avoid the clumsiness of Neville Longbottom. You should have no "lost in translation" moments between a shortcut and its normal form.

I kinda wish that the Notebook editor would have a means of flipping something to its normal form (and vice versa). I also wish that the Help documentation would always show both forms on examples; it tends to be prejudiced towards the shortcuts. I like to verbalize the normal form when I look up something in the docs.

POSTED BY: Phil Earnhardt
Posted 11 months ago

That's a fine and helpful tip for the newbies to the Language. I'm having to constantly look up and flip over to several views or points. The Docs but There are many "points" in the docs to look at. I have to reference a LOT. So that's one thing I'm learning. It's not enough just to read one part.

I often remind myself. I'm doing it wrong! What am I missing? What can I not see?

Does anyone have any experience with Reeturn[] I'm not able to use it on TraceDialog. With this error message.

"TraceDialog::dgbgn: Entering Dialog; use Return[] to exit." I didn't get a dialog and my process was stuck running. I had to close the window to get it to stop.

Cheers.

POSTED BY: Zachary Wertz
Posted 11 months ago

Will there be another one of these classes?

POSTED BY: Paul Nielan
Posted 11 months ago

Will there be another one of these classes?

POSTED BY: Paul Nielan
Posted 11 months ago

I'm having problems understanding number three which works with Set Delayed Rule and Blank.

I couldn't catch on C++ a long time ago in college. Maybe that's why I don't understand this concept. Object orientated programming. My problem seems that there's too much overlap in definitions. There's no similarly to compare it to. The concept is Invisible.

There's a few things happening with the blank, and the "f" and the "x" those are all condensed things or overlapping definitions. Left to right replacement. Got that. what is "f" and "x" or x or x you could call them "Head" or a function. of the [ ]. I'm struggling to find an explanation of what x or x is Now I do understand that is is just a blank to be filled in but the x or f is a distraction. why not just _ without f or x that make more sense.

Now I tried to make this work and nope. nothing works. Is it because of the Delayed rule? That's a new idea. I guess I haven't tested that out yet. changing :> to -> instead. I can try that next. I don't understand what to do to get the replacement happening it seems that it should be just happening if there are blanks in the right place but maybe it's because of the Delayed rule. I tried changing the delay to a rule and it's the same. I used Trace. With Trace I just get {} and without Trace I get the same out as I put in f[1,2,3] I don't know if the rule is even working or able to work on f[1,2,3] that's my problem I don't see how the problem is applied to f[1,2,3] it seems it is not.

POSTED BY: Zachary Wertz
Posted 11 months ago

I found Richard J. Gaylord's video on youtube. It's a great video and I saw it a while back. The first five minutes of the video or really the first minute of the video when he starts his lecture has got me into the invisible part and it's a great help to understand the structure of wolfram language. I hope to find and learn how to read the Docs better.

I'm doing FullForm the rhs of number three which is what I was talking about. It's been helping so far. I still have to learn the Pattern. Like I said it's confusing how the new vocabulary is used with seeming overlapping meanings. This first minute of the video has been helpful to find what to look for.

Head and what follows? and how WF evaluates it maybe old news for some of if you. For me who I haven't been using the language for a long time regularly it's a big help. I'm not done yet.

POSTED BY: Zachary Wertz
Posted 11 months ago

Thanks for the useful reference. Fills in some of the blanks in Arden's excellent lectures. I always like several points of view.

POSTED BY: Steven Brawer

This is a nice talk! Glad to see that I've naturally settled on many of the same explanations as Richard gives here :)

POSTED BY: Arben Kalziqi
Posted 11 months ago

I worked on a exercise to from EIWL which was just adding up a list and I wanted to do it with Plus, Range. I failed and spent a hour learning. Then came back to it.

Recap. I was working on trying to use a symbol to do the adding or multiplying with @@ I could not get it to work. I could not figure out what was happening. I tried a lot of things. Nothing seemed to work. I looked up Trace, TreeForm, and a whatever I came across. I played around with the documentation changing the code around.

I figured out how to do it, after my nap. Checking with TreeForm told me the "path" of the code that I could not see. Checking the documentation on how to properly write it over and over. I could see that Plus was not getting worked on the List. So I knew I wrote it wrong.

The hour of work was not wasted. From that hour of work. That is how I got the code to be "visible". I worked with TreeForm and the documentation trying out Map, ThreadMap, Funny because I was using Apply in the first place I just wrote it wrong.

I found a you-tube video on Lists that was helpful. I also read some tutorials on how to properly write wolfram code. I read the Filled in Notebooks, and whatever I came across. So it was a bunch of hopping around sources then a long break and back to finally finding the correct way to write it. It helps that I realize that I'm trying to write it the way I think it works and that is wrong. I need to find how it is supposed to be written.

Merry coding to everyone.

POSTED BY: Zachary Wertz

Hi Arben;

As an aside, I came across something the other day that I did not understand but figured you would.

I understand what a Derivative (D) and an Implicit Derivative (ImplicitD) are, but what in the world is a Total Derivative (Dt). In all my years of math, I just came across this thing called a Total Derivative and hoped you could shed some light.

Thanks,

Mitch Sandlin

POSTED BY: Mitchell Sandlin
Posted 11 months ago

Hey Mitch! I also remember coming to this comparatively late in my math career (though frankly I'm sure it was in our diff eq class and I missed it because I was boooooored). As is indicated by Steve, probably the easiest inroad to the idea of the total derivative is the chain rule. To that end, check out the corresponding lesson from our multivariable calculus MOOC (including lesson notebook, lecture video, and exercises with solutions)—it's lesson 15 here.

POSTED BY: Arben Kalziqi

On the L1 exam, do I take a list shown as {1, 2, 3, a, b, c} verbatim or am I to interpret the list as {1, 2, 3, "a", "b", "c"}? I wasn't sure since the color coding as done in Mathematica didn't carry over to the exam.

POSTED BY: David Snyder

Definitely verbatim. If you think that there is an error, be sure to tell Arben :-)

POSTED BY: Lori Johnson

(confirming here for, well, official confirmation.)

POSTED BY: Arben Kalziqi
Posted 11 months ago

I have a question about pattern matching.

  1. Cases[{1, 1, 0, 1, 1, 0, 1, 0, 0, 1, 0, 1, 0, 1, 0, 0}, {0, 1, 0}] This returns empty. I know why :
  2. Cases[{1, 1, 0, 1, 1, 0, 1, 0, 0, 1, {0, 1, 0}, 1, 0, 0}, {0, 1, 0}] This returns {0,1,0}

So the List delimiters {} make that example 1 does not work.

I tried Cases[{1, 1, 0, 1, 1, 0, 1, 0, 0, 1, 0, 1, 0, 1, 0, 0}, {_ _ , 0, 1, 0, _ _ }]
Code should show { underscore 3x , 0,1,0, underscore 3x } unfortunately but comes out different in the post (only 2 underscores are shown). This also does not give a correct result but {}. I was surprised that this is wrong :-)

How should I approach example one, so that I get the 4 times 0,1,0 as a result? And how do I get the starting positions of each 0,1,0 ? (so here : {6,9,11,13}

POSTED BY: B. Cornas
Posted 11 months ago

Use Partition. It can be done in one instruction, but to make it clearer, I break it down:

lst = {1, 1, 0, 1, 1, 0, 1, 0, 0, 1, 0, 1, 0, 1, 0, 0};

par = Partition[lst, 3, 1]

Cases[par, {0, 1, 0}]
POSTED BY: Gerald Dorfman
Posted 11 months ago

Thanks Gerald, I understand your way. I see that the '1' in the Partition[lst, 3, 1] is needed. I also need to get the positions and can get those by Position[par,{0,1,0}]. The result is correct. Your way works also good if the list starts or ends with the wanted pattern.

In a broader context, I need to test many more 'regularities'. Imagine the lst = {1, 1, 0, 1, 1, 0, 1, 0, 0, 1, 0, 1, 0, 1, 0, 0} is a rhythm where 1 stands for a hit and 0 for a silence. I want to get all the regularities from this rhythm line. So I would need to test also parts with length 2, length 4, etc. And I need to do it fast (testing many rhythms). So maybe there is a way to avoid the partitioning, which would have to be different for different lengths of search lists.

This checks for 'repetitions'. How could I test for mirror symmetry? E.g. {0,0,0,1,0,1,1,0,1}. here 1,0,1 is a symmetry as well as 1,0,1,1,0,1 How could I test for Permutations of the pattern? E.g. {1,0,1,x,x,x,x 0,1,1,x,x,x,x,1,1,0}. here are 3 permutations of {1,0,1} present. How to detect them?

POSTED BY: B. Cornas
Posted 11 months ago

In attached file Symmetries.nb , I've written and executed a module to find the symmetries you are interested in and a function to count the number of subsequences in an ordered list of length at least 2.

The module creates a random list of zeros and ones of length at least 2, finds all its ordered subsequences of length at least 2, and lists all its symmetric subsequences. I assume that subsequences of length 1 are of no interest because they are trivially symmetric.

The simple function counts the number of subsequences of length 2 or more from a list of length at least 2. The code uses the function to print a table of the counts for lists of length 2 through 15.

Although I compute the number of subsequences as a sum, it's easy to derive the formula by just evaluating the sum. Given a list (i.e., an ordered list) of length n, the number of subsequences of it of length k is just the number of leftmost positions of sequences of length k, namely n-(k-1) = (n+1) - k. Now, sum for k = 2 to n. Since (n+1) is a constant, the sum of the first term is (n-1)(n+1) = n^2 - 1. The sum of the second term is just 1 less than the sum of the first n integers, namely n(n+1)/2 - 1 = (n^2)/2 + n/2 -1. So the number of subsequences of length at least 2 of a list of length n is (n^2 - 1) - ((n^2)/2 + n/2 -1) = ((n^2)/2 - n/2 = n(n-1)/2. So, instead of using the sum to define the function numSequences , one could define it as numSequences[n_Integer /; n >= 2] := n(n-1)/2 . I used the sum just to make the code easy to understand.

Attachments:
POSTED BY: Gerald Dorfman
Posted 11 months ago

Hi Gerald, thanks for your help. I just downloaded your example Nb and will look into it. I'll come back on it later.

POSTED BY: B. Cornas
Posted 11 months ago

I worked through your Nb. Thanks for the work you've done. It works well and I learned a thing or two, which I can apply on other places in my project.

Really appreciate your help, as well as Rohit's.

POSTED BY: B. Cornas
Posted 11 months ago

Here are a couple addenda to the Symmetries.nb file and comments I provided:

(1) A quick derivation for the formula for the number of (ordered) subsequences in an n-sequence (i.e., an ordered sequence of length n): Call the n-sequence "seq", and number its positions from left to right 1, 2, ..., n. The subsequences are uniquely defined by their leftmost and rightmost positions. Since we want subsequences of length at least 2, the choices for the rightmost position are 2, ..., n. If k is the choice for the rightmost position, then the choices for the leftmost position are 1, 2, ..., k-1, so there are k-1 choices for the leftmost position and hence k-1 subsequences with rightmost position k. So, the number of subsequences in an n-sequence is the sum of (k-1) for k = 2, 3, ..., n. That's the same as the sum of the integers from 1 to (n-1), which is just n(n-1)/2. (The formula for the sum of the first m integers is well known. It's just m(m+1)/2.)

(2) Instead of using Cases in file Symmetries.nb, I could have used Select:

    Select[subsequences, (#===Reverse[#])&];

I used Cases to show how to use a pattern, which can also be used for Position.

POSTED BY: Gerald Dorfman
Posted 11 months ago

I am happy to see your explanation for the derivation for the formula for the number of (ordered) subsequences in an n-sequence. Very helpful. I tried Select i.s.o Cases and that works fine as well. I guess that you put the parenthesis around # === Reverse[#] in Select[subSequences, (# === Reverse[#]) &]; for clarity. It also works without them.

Thanks again , Gerald

POSTED BY: B. Cornas
Posted 11 months ago

Yes. I put the parentheses around the pure function for clarity. Doing stuff like that was recommended in class. For myself, given how simple that pure function is, I would not have bothered.

POSTED BY: Gerald Dorfman
Posted 11 months ago

I thought so, but a useful practice :-)

POSTED BY: B. Cornas

one quick note here: parentheses are still good for clarity, but we did—in version 13.x for some value of x—introduce something to help here. namely, when you type & after a pure function, it will quickly highlight the function it's "closing" so you can be sure that you have things grouped correctly. (the common case here is when doing something like ColorFunction->#1^2&—you need the parentheses after the -> because otherwise it thinks the pure function to be closed is ColorFunction->#1^2 rather than just #1^2.)

POSTED BY: Arben Kalziqi
Posted 11 months ago

I have not run any benchmarks but I suspect that using the built-in PalindromeQ may be faster than # === Reverse[#] &.

POSTED BY: Rohit Namjoshi
Posted 11 months ago

Hi Rohit,

that was my idea as well and I wanted to do some timing tests as soon as I had time. Also the Option 'Overlap' we talked about earlier is very useful for my project.

But the extended explanation of Gerald helped me in another way. I am pretty new to Wolfram language and still have to adapt from Procedural programming, which takes quite a paradigm shift.

POSTED BY: B. Cornas
Posted 11 months ago

Use the Sequence* family of functions

list = {1, 1, 0, 1, 1, 0, 1, 0, 0, 1, 0, 1, 0, 1, 0, 0}

SequenceCases[list, {0, 1, 0}, Overlaps -> True]

SequencePosition[list, {0, 1, 0}]

Overlaps -> True is the default for SequencePosition which is a little inconsistent

POSTED BY: Rohit Namjoshi
Posted 11 months ago

Great Rohit, this works very well. Thanks.

In which sense is Overlaps -> True inconsistent? I saw in the docs, that SequencePosition has the same Overlaps option. Is it also inconsistent there?

Would you also have some suggestions for the following :

This checks for 'repetitions'. How could I test for mirror symmetry? E.g. {0,0,0,1,0,1,1,0,1}. here 1,0,1 is a symmetry as well as 1,0,1,1,0,1 How could I test for Permutations of the pattern? E.g. {1,0,1,x,x,x,x 0,1,1,x,x,x,x,1,1,0}. here are 3 permutations of {1,0,1} present. How to detect them?

POSTED BY: B. Cornas
Posted 11 months ago

Hi B.

In which sense is Overlaps -> True inconsistent? I saw in the docs, that SequencePosition has the same Overlaps option. Is it also inconsistent there?

What I meant is that the default value of Overlaps is True for SequencePosition and False for SequenceCases. That seems inconsistent to me.

For symmetry something like this?

list = {0, 0, 0, 1, 0, 1, 1, 0, 1}
Subsequences[list, {3, Length@list}] // Select[PalindromeQ]

For permutations generate all of them and test each with SequencePosition?

perms = Permutations[{1, 0, 1}]
list = {1,0,1,x,x,x,x 0,1,1,x,x,x,x,1,1,0}
SequencePosition[list, #] & /@ perms
POSTED BY: Updating Name
Posted 11 months ago

Thanks Rohit, i get the 'inconsistency' and I agree.

Your solutions for Symmetry and Permutations work fine and I can extend on them :-)

POSTED BY: B. Cornas
Posted 11 months ago

Hi Rohit,

I wanted to take your idea a little further, more general, but I run into a snag I do not understand. SequenceCases[pattern1, {0, 1, 0}, Overlaps -> False] from your previous post. This works great :-)

But now to find more general patterns : list={0, 0, 1, 1, 1, 1, 0, 1, 0, 1, 1, 0, 1, 0, 0, 1} patternToFind = {x, y, x_} SequenceCases[list, patternToFind, Overlaps -> False]

I get {{1, 1, 1}, {1, 0, 1}, {1, 0, 1}} as result.

Why is the [1,1,1} there? I specified the first and third to be the same, but the second element to be different. I expected to get {1,0,1} and {0,1,0}. The result also misses the {0,1,0}.

Any good ideas?

POSTED BY: Updating Name
Posted 11 months ago

Hi B.

In a pattern with a symbol, the symbol is what the match is set to. So {x_, y_, x_} matches something, x is set to that thing, followed by something, y is set to that thing, etc. The fact that the symbol is different does not mean the match has to be different.

MatchQ[{1, 1, 1}, {a_, b_, a_}]
(* True *)

You need to use Condition to ensure they are different

MatchQ[{1, 1, 1}, {a_, b_, a_} /; b != a]
(* False *)

The missing ones are because of Overlaps -> False

SequenceCases[list, {x_, y_, x_} /; x != y, Overlaps -> True]
(* {{1, 0, 1}, {0, 1, 0}, {1, 0, 1}, {1, 0, 1}, {0, 1, 0}} *)

The matching positions

SequencePosition[list, {x_, y_, x_} /; x != y, Overlaps -> True]
(* {{6, 8}, {7, 9}, {8, 10}, {11, 13}, {12, 14}} *)
POSTED BY: Rohit Namjoshi
Posted 11 months ago

Great Rohit, thanks.

I had understood that {x,y,x} was different from {x,x,x} as a pattern .

I had the 'Overlaps -> True' first, but as I did not have x != y , I got about way too much solutions.

Now it works fine :-)

Best , Bert

POSTED BY: B. Cornas
Posted 11 months ago

Hi Bert,

They are not the same

MatchQ[{1, 2, 1}, {x_, x_, x_}]
(* False *)

MatchQ[{1, 2, 1}, {x_, y_, x_}]
(* True *)

The first pattern specifies that all elements must be the same. The second specifies that the first and third elements are the same. The second element can be anything, including the same as the first and third.

MatchQ[{1, 1, 1}, {x_, y_, x_}]
(* True *)
POSTED BY: Rohit Namjoshi

Precisely! When I talk about pattern matching, I always talk about x_ as "something—call it x" for exactly this reason.

POSTED BY: Arben Kalziqi
Posted 11 months ago

Rohit & Arben, I get it. Great help. Cheers , Bert

POSTED BY: B. Cornas
Posted 11 months ago

Is the 5th question in the level 1 exam :-

"Beginning with data = {{15000, , 89000,, ............"

correct/valid ?

POSTED BY: Doug Beveridge
Posted 11 months ago

Ok , it is correct

POSTED BY: Doug Beveridge
Posted 11 months ago

Questions are randomly (actually, pseudorandomly) selected from a pool; the odds that any person would have the same 5th question are rather low.

I had a question on my L1 quiz where there were 3 possible answers, plus "all of the above". It turned out that 2 of the 3 were giving the correct response and one was giving a different response; none of the answers could possibly be right. I was running my quiz notebook off of the Wolfram Cloud; that appears to be the reason for the one buggy response. I answered the question as if the function was behaving correctly. I sent an e-mail to wolfram-u@wolfram.com noting the bad behavior and filed it as a bug in Mathematica.

Here are my rules in taking a Wolfram quiz or cert test. It took me a while to figure these out...

  1. Try as hard as you can on every single question.

  2. Realize that you may still get some questions wrong no matter what you do or know. Don't sweat it.

  3. Appreciate the difficulty in authoring a perfect quiz. There may be bugs, bugs may creep in over time, and you may get some "no win" questions. Flag questions you're pretty sure are buggy in e-mail to Wolfram U. Remember rules 1 and 2, and have fun.

POSTED BY: Phil Earnhardt

Thanks Phil—no kidding on authoring assessments! It's difficult to write questions of the correct difficulty, the right amount of "almost, but not exactly, and wrong in an instructive way" answers, and account for possible bugs, input types, edge cases, and so on. Still, we're always looking to improve and always appreciate any bug reports or suggestions that come our way.

POSTED BY: Arben Kalziqi
Posted 11 months ago

Hi Arben,

I have another question about Manipulate and it ties in with my questions from 3 days ago about 'Stepping back' to previous settings in Manipulate.

One has to be quite careful not to create a loop in a Manipulate construct. E,g. a= a+2 gives already a loop. If done like aNew = a+ 2, then it's ok. More 'funny things' might happen in a more complex Manipulate construction.

My question is : How is the order of the calculations in a Manipulate? Does it go from top to bottom, like line for line (and left to right)? Always? What if I change a parameter (slider) whose appearance is only after 20 lines of code? Is still all the code run from the top, or does the re-calculation start where the changed value first appears (e.g. line 20)?

How can I step through the code and watch the changing variables, like in 'Step-Debug'? Thanks a lot,

Bert

POSTED BY: B. Cornas
Posted 11 months ago

Hi Arben,

I would really appreciate if you could shed some light on these questions of mine. I hope you find the time for it :-)

I read through both "Dynamic & Manipulate Notes" and the "Advanced Dynamic & Manipulate Notes". I could not find anything about the order of execution of the lines of code or step Debug in a Manipulate construct.

Thanks a lot, Beat

POSTED BY: B. Cornas

Hey Bert! I didn't know the answer to this one off the top of my head. After some investigation, I have to say "it's complicated."

From the Advanced Dynamic Functionality Tutorial:

Dynamic expressions can be nested, and the system takes great care to update them only when necessary. Particularly when the contents of a Dynamic contain further interactive elements, it is important to keep track of what will stay static and what will update, when a given variable is changed.

It seems as if for any given dynamic expression, there's a step which involves "tracking down" what pieces of the expression can/will change when certain dynamic variables are updated, and it's those pieces which end up being reevaluated.

POSTED BY: Arben Kalziqi
Posted 11 months ago

Thanks Arben, for your help and work.

It certainly is quite complicated and I started wondering after I found that the code inside Manipulate is not necessarily executed from top to bottom. It is getting even more complicated if there are Dynamic functions explicit withing the Manipulate. Maybe I can come up with some Flags that get set when some line of code is re-executed. I'll have to try.

I will read through the Advanced manipulate and Dynamic paper again :-)

Best, Bert

POSTED BY: B. Cornas

Part of the issue here—and I think why you probably didn't think that the quoted section was useful when you first read it—is that Manipulate is making its own DynamicModule under the hood. This means that even if you're not typing Dynamic[x], the fact that you have (say) {x,1,10} as the second argument to Manipulate means that the constructed DynamicModule does have Dynamic[x] in there—which is why the quoted section is actually relevant, even if you aren't personally typing Dynamic[x].

POSTED BY: Arben Kalziqi
Posted 11 months ago

Very good point, Arben :-)

POSTED BY: B. Cornas
Posted 11 months ago

I am having email problems. How do I join the class without today's link? Sorry about this!

POSTED BY: Paul Tikotin

@Paul Tikotin, we'll send your join link to your email, or you can contact us at wolfram-u@wolfram.com to provide a different address, if you're having trouble with gmail. You can also go to the webinar page to register and receive a join link using a different email address. Email communication of some sort will be needed.

POSTED BY: Jamie Peterson
Posted 11 months ago

I think I have just sorted the problem,...Thank you!

POSTED BY: Paul Tikotin
Posted 11 months ago

I am running Mathematica version 14 under Windows 11 on an HP Envy laptop. To compare AbsoluteTiming for 3 alternative instructions to do the same thing, I wrote the following code and was surprised to see that "Power" was not quite the fastest. (Timing[...] just showed 0. for all 3 instructions.)
I'd appreciate an analysis of my findings.

In[104]:= listOfAssoc=Association@@@({"a"->#[[1]],"b"->#[[2]]}&/@Partition[Range[1000],2]);
tbl=listOfAssoc[[;;2]]

Out[105]= {<|a->1,b->2|>,<|a->3,b->4|>}

In[106]:= First@AbsoluteTiming[Power@@@listOfAssoc]

First@AbsoluteTiming[#a^#b&/@listOfAssoc]

First@AbsoluteTiming[#[[1]]^#[[2]]&/@listOfAssoc]

Out[106]= 0.0118669

Out[107]= 0.0107371

Out[108]= 0.0082883
POSTED BY: Gerald Dorfman

As we're fleshing out project ideas for the level 2 certification, is there someone we can meet with to make sure we're on the right track?

POSTED BY: Kari Grafton
Posted 11 months ago

How about a separate Wolfram Community discussion to speculate/brainstorm about L2 cert projects? Anyone can create a new discussion, but it might be good to have the "seal of approval" if @Arben Kalziqi or @Jamie Peterson were to create it. :)

I am looking at a Wolfram Examples project showing muscle co-activation of hand flexors/extensors, arm flexors/extensors, and arm pronators/supinators. The beautiful Anatomy package has the ability to do color illustrations; I can graphically show the muscles involved in particular co-activations. Apparently, the Anatomy package also has identifiers for the antagonist pairs; I need to investigate that. Apologies if those words make no sense; it's better in pictures.

POSTED BY: Phil Earnhardt

I really like this idea! @Arben Kalziqi, @Jamie Peterson, can we make this happen?

POSTED BY: Kari Grafton

While I (hopefully clearly) do enjoy answering everybody's questions here, for something as involved as the Level 2 projects I'm afraid it's not really feasible from a time/involvement perspective. Like Jamie said, please do take advantage of office hours, and we may well offer more things like this in the future. (Actually, typing that reminds me that we have something in the cards for like, an "interesting example + office hour Fridays" series... we'll see if that comes to fruition!)

POSTED BY: Arben Kalziqi
Posted 11 months ago

While I (hopefully clearly) do enjoy answering everybody's questions here, for something as involved as the Level 2 projects I'm afraid it's not really feasible from a time/involvement perspective.

Whoa! While I can't speak for @Kari Grafton, my intent was never to turn this proposed discussion group into a place for Wolfram staffers to guide us through any steps of the L2 project process. My intention was to have an informal conversation between participants in the process. In what I'm visualizing, it's probably the case that WR staff participation in the discussion would be counterproductive.

As proposed, the L2 cert process is a solitary interaction between a developer and Wolfram Research. The process is isolated; there's no sense of community in that process. A successful course (like this one) nurtures a sense of community in the participants; that's what I was looking for in a L2 Wolfram Community discussion.

Maybe the discussion group is about encouraging participants to create a project worthy of putting in one of the designated repositories. That project may or may not be sufficient for a L2 certification; that's not really important. I think this is very different from what Arben and Jamie were hearing when they reacted to this discussion.

What is clear: I need to choose words carefully for the first message of the discussion group. Let people know this has nothing to do with anything official that WR is doing. Let them know it's about community. The discussion is more like Stephen Covey's idea in The 8th Habit: Find Your Voice, and Inspire Others to Find Theirs

POSTED BY: Phil Earnhardt

In that case, I support y'all wholeheartedly (imagine I could use emojis here and added a :sweat_smile: at the end there, okay).

POSTED BY: Arben Kalziqi

My best suggestion is to use today's Study Group survey to request your registration at the March 1 instructor office hours, which will also be available as part of the upcoming Programming Proficiency course sequence. Register for the course sequence here.

POSTED BY: Jamie Peterson

Great, thank you! I did request registration for the office hours. And I agree that a L2 discussion board geared towards the participants is a great idea.

POSTED BY: Kari Grafton

Hi Jamie, Cassidy, Abrita, Arben!

Thank you for all your hard work that goes into making these study groups enjoyable. [big smiley]

I enjoyed the fill-in-the-blank format of the Level 1 quiz. Keep that coming. The Level 2 Certificate is on my agenda (the Nutrition Report looks interesting) as are almost every new webinar and WSG. Artisinal Calculus, why, yes, I think I will!

I'm sleeping in today but I will see y'all soon. Stay safe. Stay healthy. Lori

POSTED BY: Lori Johnson

Same! I can't get enough of these webinars and try to do every one that looks like it's even remotely related to something I might use. Or just looks particularly interesting.

POSTED BY: Kari Grafton

Hi Kari!

Nice to meet you! Absolutely. My goal is to learn at least one new piece of creative programming each time, be reminded of a command I may have forgotten, etc. I am seldom disappointed :D

Do you collect session notebooks, too? They're like conference handouts. There are always "goodies" inside.

POSTED BY: Lori Johnson

Thanks Lori! We always enjoy having you with us. Hopefully you enjoy the review recording and notebook if you'd like to peruse them after sleeping in!

POSTED BY: Arben Kalziqi
Posted 11 months ago

Aww, that's a very nice thing to say. Thank you, Arben!!!

Will do! Have a fun weekend :D

POSTED BY: Updating Name

I am trying to rid my code of Do loops as suggested in the training, but I am having some difficulty in getting the simple example in my notebook to work. It is not clear to me how to get the names to cycle through the example. I have several sources of data each with a key value that I am trying to combine into one variable. This is how I am doing it now. I believe if I could figure out how to do the simple example in my attached notebook, I could make the below work.

Do[bE1defgrad[[i]] = 
   Map[Append[#, Select[sEDNH, MemberQ[#, bE1[[i]]] &][[1]][[3]]] &, 
    Map[Append[#, 
       Select[pressure, MemberQ[#, bE1[[i]]] &][[1]][[2]]] &, 
     Select[defGrad, MemberQ[#, bE1[[i]]] &]]];,
 {i, 1, bE1len}]

POSTED BY: Philip Van Riper
Posted 11 months ago

Hi Philip,

The code you posted is incomplete, there are several symbols that are not defined sEDNH, pressure, defGrad, ....

I also don't understand what you are trying to accomplish with combined. It just adds an additional level of nesting to dgrad because names includes all of the elements in dgrad

dgrad == First /@ combined
(* True *)

If you just want to pick out the matches to names here is a functional way to do it. There is no need to initialize combined

combined = Select[dgrad, MemberQ[names, First@#] &]
POSTED BY: Rohit Namjoshi

Hey Philip—agree with Rohit here; I'm not sure what you're trying to do exactly. If you could upload a version of the notebook where you've defined everything that you're using and where you've explained what each "step" is supposed to do (like by saying "i have [input], and it does [xyz], resulting in [output]"), I'm sure I'd be able to help.

POSTED BY: Arben Kalziqi

Thanks for the feedback. I am reading several files into my notebook and I ended up simplifying it too much without providing the needed context. Hopefully the attached notebook rectifies this. I basically have three lists that contain data keyed to a list of names. I wish to combine all of the data into one list, but only for the items in the list names.

My background is Fortran, so I instantly think of loops, however the Wolfram language intrigues me and I really want to use it effectively. Phil

POSTED BY: Philip Van Riper
Posted 11 months ago

Hi Philip,

I have modified your notebook and added a solution to the end. Light green cells are ones I added.

Attachments:
POSTED BY: Rohit Namjoshi

Rohit, Thanks for the solution. It looks much better that the Do loop and gives me new avenues to explore. Phil

POSTED BY: Philip Van Riper
Posted 11 months ago

package question

All examples of creating a package (after foo::usage=...;bar::usage=...., etc) use

Begin["`Private`"]

and then function bodies follow. Does the word "Private" have some particular significance. What would be the problem with using some other word like

Begin["`NotPrivateAtAll`"]

How is this designation actually used?

POSTED BY: Steven Brawer
Posted 11 months ago

Does the word "Private" have some particular significance. What would be the problem with using some other word like [...]

"Private" has no significance. If you go to the Wolfram Language Documentation for Begin, you'll see the examples there use "MyContext`". "Private" is not a keyword in the WL, but [IMHO] the fact that it is a keyword in many programming languages means it should probably be avoided in examples in Wolfram courses. OTOH, "MyContext" is definitely un-keywordy.

It's really helpful to use the Wolfram Language Documentation as the first place to go with questions like this. I don't think that Arben has emphasized the wonderful Wolfram documentation in this course. Docs are available on the web; they are also typically installed in a Mathematica install and can be accessed through a Mathematica window. Carpe documenta!

POSTED BY: Phil Earnhardt
Posted 11 months ago

Thank you. Mathematica documentation could be called "minimal".

Consider the "Begin" documentation. For example, under "applications" there is this for BeginPackage: "Make symbols used for package function definitions private, reducing the possibility for conflict:" Fine. What exactly is the problem with saying: "this defines a namespace" (or at east say Context = namespace)? Unless it doesn't actually define a namespace (or it does, but with lots of caveats), and if not, that should definitely be explained.

Moreover, "Private" is used in the one single "application" section example but not in any examples of the "examples" section. In addition, without a specific statement that "Private" is not special, it is unclear whether using it might have some unanticipated side effect. (In Mathematica, unanticipated side effects are anticipated.) Absence of a comment is not documentation, especially these days when everyone is very busy. If you google "Mathematica packages", there is hardly anything there (or hardly anything simple).

Mathematica is a wonderful and very useful but very complicated system, and the paltry number of examples for such an important subject as Packages means hours wasted trying to figure simple things out.

Here's another interesting thing that is hardly obvious (at least to me, and is probably buried somewhere in a document). Suppose you define G = {x1^n, x2^m} in a program (x1,x2, unassigned symbols) and you want to differentiate G (with respect to x1) in a function someName[G] := ... which is located in a package, where the function body contains D[G[[1]],x1]. As far as I can tell, unless you define the package function as someName[G, x1,x2] and call it from the program as someName[G,x1,x2], it will NOT work. How does the engine know to link the x1,x2 in the call with the x1,x2 in G? This is hardly obvious. There is presumably a whole world of Mathematica-namespace concepts underneath this, which it would be very nice to at least glimpse.

POSTED BY: Steven Brawer
Posted 11 months ago

If there are bugs/shortcomings in the documentation, you can file a support ticket with Wolfram.

As far as Private not being a keyword, I just looked up that word in the documentation. Simple. "Private" the String is used as an option in some functions, but that doesn't leak out and affect other things. If a word isn't mentioned in the doc entry for some function, it shouldn't have any magic way to affect that function.

The other thing I noticed with your questions: if you're curious what happens when you do X, just try it. If the kernel behaves in some inexplicable way, that's the point in time to investigate further. I imagine Stephen Wolfram sitting (or standing) in his office in MA; I bet it's in his muscle memory just to try something about before looking it up in the docs or (gasp!) asking another staffer for clarification.

POSTED BY: Phil Earnhardt
Posted 11 months ago

Is there any difference between Proficiency in Wolfram Language Level 1 Certification and Proficiency in Mathematica Level 1 Certification ? It seems that the requirements and the process are the same.

POSTED BY: Artur Piekosz

Good question! While there is significant overlap between the Wolfram Language and Mathematica Level 1 certifications, we are making changes to focus the Wolfram Language certification on programming, and the Mathematica certification on using the product for technical computation. The Level 1 pages and the exam questions will be reflecting these latest changes in upcoming days and weeks.

POSTED BY: Jamie Peterson

Hi there!

I have a question about adding a new variable (Key-> Value) to a list of associations. For instance I recreate the following data

{V1, V2} = {RandomVariate[NormalDistribution[15, 4], 50], 
RandomVariate[NormalDistribution[20, 7], 50]}; (*Create two random variables*)
data = AssociationThread[{"Diameter", "Height"} -> #] & /@  Thread[{V1, V2}];
(*Arrange a list of association*)

Suppose I want to add a new variable, for instance, the logarithm of Diameter variable, however, the only way I succed was using a implicit loop using the table function:

V3 = Log[data[[All, "Diameter"]]];
Table[Append[data[[i, All]], "LogDiameter" -> V3[[i]]], {i, 1, 
Length[V3]}];

I would like to ask if I could add a variable to the association list avoiding the Table function?

Thanks in advance!

Posted 11 months ago

Hi Andres,

Here is one way to do it

augmentedData = <|#, "LogDiameter" -> Log[#Diameter]|> & /@ data
Dataset@augmentedData

This relies on the implicit addition of key/value pairs to an Association when wrapped in Association. E.g.

assoc = <|"a" -> 1, "b" -> 2|>
<|assoc, "c" -> 3|>
(* <|"a" -> 1, "b" -> 2, "c" -> 3|> *)
POSTED BY: Rohit Namjoshi

Hi Andres—here are a few implementations that should get your thoughts moving,

If you have access to V1 and V2 directly, you can do:

new = AssociationThread[{"Diameter", "Height", "LogDiameter"} -> #] & /@ 
   Transpose[{V1, V2, V3}];

If you only have direct access to data, you could do this (honestly, I'd do something like this anyway):

new = <|"Diameter" -> #Diameter, "Height" -> #Height, 
     "LogDiameter" -> Log@#Diameter|> & /@ data;

The fastest and shortest way has got to be something like:

new = Append[#, "LogDiameter" -> Log@#Diameter] & /@ data;
POSTED BY: Arben Kalziqi

I just wanted everyone to know that the notebooks for the book "Query" are available at: Query notebooks. I have found them to be a helpful companion to our class.

POSTED BY: Michael O'Connor
Posted 11 months ago

When I click on the Grading Rubric link on this page I get

HTTP Error Code 403 Sorry, you do not have permission to access this item.

POSTED BY: Rohit Namjoshi

Hi Rohit, this should be fixed now—works for me, as well. You may need to clear cookies/cache to get it to work, though, as a version with the wrong permissions may be cached on your end.

POSTED BY: Arben Kalziqi

How to create the association command, "<|", vertical bar?

POSTED BY: Taiboo Song

It will depend on your keyboard layout, but a QWERTY layout should have the | accessible by pressing shift and . The < is just a less-than sign, so that's <|association here|>.

POSTED BY: Arben Kalziqi
Posted 11 months ago

Arben

The backslash needs to be escaped in the reply window. Shift \

Shift + \

POSTED BY: Rohit Namjoshi

oh god, it escapes even when there's nothing in front of it? blegh, thanks Rohit!

Taiboo: indeed, you want shift + backslash—shift + \

POSTED BY: Arben Kalziqi

Hello Arben: You are an excellent instructor and learning a lot.

When I open your filled-out notebook file, the file shows up as the notebook, with nice colors. But when I open the file class file, it looks different. Why is that and how to get into a nice notebook format for your class notes?

POSTED BY: Taiboo Song

Hey Taiboo—we have two versions of our notebooks, generally speaking. One is the instructor version, which I use when teaching live and which we use for making edits. Then, there's a more monochrome "handout" version which has some formatting changed for consistency's sake and to work nicely on the cloud.

POSTED BY: Arben Kalziqi
Posted 11 months ago

Arben - Two questions:

1.    On the Quiz, is there a way to find out which problem(s) we got wrong?

2.    My "Day01-ListsAndAssociations-Filled.nb" is filled with NULLs. Is there a way to resolve this?

Thanks in advance,

POSTED BY: Updating Name
  1. I don't think so, but if you have a question about a particular question you can ask me. (I also think that if you email us, we can manually pull your answers and which were correct.)
  2. No Nulls for me. You might get them if you don't Enable Dynamics when prompted.
POSTED BY: Arben Kalziqi

The quizzes used to indicate which ones were incorrect, but within the last year that was changed. I found the old method of indicating which answers were correct educational and the new method not so much.

Thanks,

Mitch Sandlin

POSTED BY: Mitchell Sandlin
Posted 11 months ago

I have some questions about the Manipulate construction: I have a manipulate with let's say 4 adjustable parameters.

  1. How can I make a button (Previous) to go to the previous values, without getting into an endless loop, which can easily happen in a Manipulate, as if one parameter is changed, everything is recalculated. I tried several ways, like putting the values in a nested list, but somehow there is always something that is not working. Ideally, I want to be able to step back several steps and also forward again (up to the last values I had). I hope my explanation is clear enough :-)

  2. How can I keep the last used values for the 4 parameters when I run the manipulate again? I believe that in a DynamicModule construction, that goes by itself.

  3. How can I easily save the values of the 4 params as a named Preset? Eg.: Preset1. I know that I can do 'Paste snapshot, which gives me all of the original code with the values of the moment, along with the result. This takes a lot of Notebook space, especially when the Manipulate code is substantial. I only want the values of the parameters, so I can load them back in later.

POSTED BY: B. Cornas

Great questions, B—any question I can't immediately answer must be pretty great ;). Let's see...

  1. I think I see what you mean, but some clarification wouldn't hurt. Could you provide a working example of a Manipulate and the behavior you'd like for it to exhibit? The thing about Manipulate is that—unless you specify otherwise via TrackedSymbols, Refresh, and similar—it will simply check for any symbol updates and it will re-compute the result. There's no way that's clear to me to get around this; the thing you're manipulating could depend arbitrarily on these inputs and other things, so it sort of needs to be that way. That said, sometimes you have a Manipulate where there's a "constant" piece which is expensive to compute, and the "dynamic" piece lives on top of it and is generally faster to compute. I don't think that that's what you're talking about, but if so, you can use With and Show to great effect for responsiveness—see attached.

  2. I have a kludgy solution for this and your third question. There's nothing stopping you from defining symbols within a Manipulate, so inside your manipulate you could say something like f[bValue=b,...] and then bValue would be Set to whatever the last value you used on your slider (or other interface element). You can use this as the starting point for the parameter too, so you'd simultaneously save all of your symbols and also be able to keep them as the starting point. Here's an example to get you going:

kValue = 1;
Manipulate[
 Plot[Sin[(kValue = k) x], {x, 0, 2 \[Pi]}], {{k, kValue}, .5, 3}]
Dynamic@kValue
POSTED BY: Arben Kalziqi
Posted 11 months ago

Somehow my post got posted under another name, so here I go again ;

Thanks Arben for your answers and for the truly excellent teaching you do. You bring something extra to the table, sometimes very subtle. Your teachings are top on all levels :-)

For the moment I reply to your answer to my questions 2+3.

I had also come up with creating secondary variables for the adjustable parameters in the Manipulate and assign the these params to the secondary variables. This way They are remembered outside the Manipulate loop. Also handy for debugging.

Building on your example, I came up with the following, which keeps the last used value when I run Manipulate again. (see code at end).

I can also make presets this way and via a Button load them back into Manipulate. I had hoped that there might be an easier way (a Wolfram function) that would do the job, as sometimes I have quite a lot of adjustable parameters in my Manipulates. But it is doable this way.

I'll post a Nb about my first question later.

Manipulate[ kValue = k; Plot[Sin[k x], {x, 0, 2 [Pi]}], {{k, kValue}, .5, 3, Appearance -> "Labeled"}]
POSTED BY: B. Cornas
Posted 11 months ago

Hi Arben, I've come up with an example Nb. It covers some of the 'Previous' and 'Next' possibilities, but not yet all. And it keeps the values last used.

Run the first cell before running the Manipulate, in order to provide initial values for the parameters inside Manipulate.

What it does is : 1. It keeps it's last values when running the Manipulate again, except for the Random pointsXY, which are newly generated each time the Manipulate is run.

  1. It steps back and forward (Button 'Previous' and 'Next'). This only works so far for the position of the lines (pointXY). Click 'Initialize X' and /or 'Initialize Y' buttons a couple of times and you can step back and forward. I limited the stepping back and forward , so no errors can occur (like asking for the a step before the first or beyond the last step).

Now I will try to be able to step back also for the other parameters, but something does not pan out yet.

Attachments:
POSTED BY: B. Cornas
Posted 11 months ago

TOPICS

It would be great if the following topics are covered, OR if there are other courses which cover them (I appreciate that some of these were touched on a bit):

custom packages , non-trivial function bodies, ordinary looping (includingPython-type looping), IDEs which work well, custom types (as in OO programming), name spaces, enums.

At least indicate what Wolfram language calls these structures, and where they are.

For example, list indexing was covered to excess, but there was only just brief touches of defining complicated functions. Custom packages are crucial for organizing any but the simplest programs, and ditto for IDEs. Sometimes speed is not important but flexibility in looping might be needed. Etc.

Thank you

POSTED BY: Steven Brawer

Thanks for the suggestions, Steven—we'll cover some of this in the coming days, and the full course offers a bit more insight than the slightly condensed study group. Once the full course is out later this month, please feel free to attend and tell us directly what you thought, if you'd like!

POSTED BY: Arben Kalziqi

Not in the file provided.

POSTED BY: Taiboo Song

Hi Taiboo, the amoeba folder where all of the materials live should have a data folder now which has all of the data used for today's presentation—as well as tomorrow's.

POSTED BY: Arben Kalziqi
Posted 11 months ago

The 2D slider

Manipulate[coordinate, {{coordinate, {0, 0},   "Coordinate"}, {-1, -1}, {1, 1}}] (* 2D slider *)

How would I do a 3D sider?

Manipulate[coordinate, {{coordinate, {0, 0,0},   "Coordinate"}, {-1, -1,-1}, {1, 1,1}}] (* 3D slider *) 

does not work

POSTED BY: Doug Beveridge
Posted 11 months ago

Doug, it would be great to have a 3D Slider. Unfortunately that is not possible on a 2D screen. How would you want to adjust the 3rd (z) axis on a 2D surface? I have been thinking more than 15 years about this and only have come up with complicated solutions, which are not truly 3D.

POSTED BY: B. Cornas
Posted 11 months ago

We Create 3D graphics on a 2D surface so we should be able to create a 3D slider .

I have noticed recently how graphics programs use a 3 button mouse to allow 3D movement in a 3D graphic , Is this not a possibility

POSTED BY: Doug Beveridge
Posted 11 months ago

We create 3D images on a 2D surface by means of perspective. To actually control 3 independent axes, you'll need 3 independent 'movements'. Of course, with extra buttons or scroll-wheels, you can add 3 third dimension, but never with a onscreen 3D slider.

What you could do, is assign the z direction to e.g. the y direction when ta button is pressed. But again, this would not be a real 3D slider, more of a work around.

POSTED BY: B. Cornas
Posted 11 months ago

Doug, it would be great to have a 3D Slider. Unfortunately that is not possible on a 2D screen. How would you want to adjust the 3rd (z) axis on a 2D surface? I have been thinking more than 15 years about this and only have come up with complicated solutions, which are not truly 3D.

If definitely possible on a 3D surface -- like what the Apple Vision Pro headset provides. That's why Apple is making such a big noise WRT the concept of "spatial computing". Check out the guided tour video published by Apple to get a taste of this. I suspect the hands-on demo -- now available at Apple stores -- would give a much better idea what this is all about. I fondly hope the Wolfram Research brain trust is checking out this tech and prototyping how they can use this tech for brilliant Spatial Computing visualizations.

On the 3D front, there is the work by Joan Horvath and Rich Cameron who have 3D printed objects designed to teach mathematical concepts: Make: Calculus, Make: Geometry, and Make: Trigonometry. AFAIK, nobody has tried to do create physical sliders in their 3DP objects; that's certainly doable. If you're really interested in 3D visualizations, I strongly suggest checking out these books and 3DP the objects that those books use.

POSTED BY: Phil Earnhardt

I also wanted to address Artur's question about @ vs. @* briefly.

It's true that in some contexts they are the same:

f@*g@x === f@g@x === f[g[x]]

but this isn't really the point of Composition/@*. The real use of this (in my experience) is chaining functions together for use in Map or Apply. Consider a basic example:

In[716]:= N@*Mean /@ {{1, 2, 4}, {4, 8, 9}}

Out[716]= {2.33333, 7.}

Here, the @* lets you write a "single composite object" which is the composition of N and Mean. This composition function is then what is mapped across the lists. Simply using @ doesn't do this:

In[715]:= N@Mean /@ {{1, 2, 4}, {4, 8, 9}}

Out[715]= {7/3, 7}

Why? What's happening—visible through Trace, which I'm going to check after I type this side note to confirm my Vibes (they were correct)—is that N@Mean gets evaluated first, and N acting on Mean just provides Mean because there's nothing there to make numeric.

So the idea is:

N@Mean /@ {{1, 2, 4}, {4, 8, 9}} = N[Mean] /@ {{1, 2, 4}, {4, 8, 9}} === Mean /@ {{1, 2, 4}, {4, 8, 9}} === {Mean@{1, 2, 4j}, Mean@{4, 8, 9}} 

whereas

N@*Mean /@ {{1, 2, 4}, {4, 8, 9}} === {N@*Mean@{1, 2, 4j}], N@Mean@{4, 8, 9}} === {N@Mean@{1, 2, 4j}], N@Mean@{4, 8, 9}} 

When composing functions, that's what you want. It can be really useful! Take the example from Good Coding Practices:

Framed@*Highlighted /@ {1, 2, 3} (* Unclear *)

(Framed@*Highlighted) /@ {1, 2, 3} (* Clearer *)

Map[Framed[Highlighted[#]] &, {1, 2, 3}] (* Clearest *)

You could rightfully argue that the first one is unclear, but if you understand how @* works it's very clear and very concise and very easy to write. You're taking the composed operation of "highlighting and then framing" and mapping it across the list {1,2,3}. It doesn't get better than that!

POSTED BY: Arben Kalziqi

The email announcing the recording availability didn't send me to the recording, just to the materials folder (no video there). Also, there was a link to the Quiz, which I'm hesitant to try without watching what went on today. Help?

POSTED BY: David Snyder
Posted 11 months ago

The e-mail announcing the availability of the recording of today's session was sent out at ~15:50 Eastern Standard Time today by the "Wolfram U Team". The subject is: "Recording available: Interfaces and Deployment". Search for "Wolfram U Team" in your inbox. The recording is accessed by clicking the "Watch The Recording" button in that e-mail. I called up the video; it is indeed today's recording.

--phil

POSTED BY: Phil Earnhardt
Posted 11 months ago

I did NOT receive this email.

POSTED BY: Steven Brawer

Please contact our staff at wolfram-u@wolfram.com if you need assistance with webinar email notifications from bigmarker.com. This is the mechanism that is used to send Daily Study Group reminders and recording notifications.

POSTED BY: Jamie Peterson
Posted 11 months ago

Jamie, it would be cool if Wolfram Research could maintain a list of those BigMarker video archive links for each user in a particular WSG and have them accessible through a Wolfram-U portal in the Wolfram Cloud. All of the user's information for that particular WSG would be stored there. That would provide a reliable mechanism for all users to have the URLs -- and not worry about losing an e-mail through spam filters, force majeure, or whatever. It would also function as a beautiful professional archive for students to reference those training materials in the future. It would be an engaging example for how users can set up small programs establishing highly accessible live data using the Wolfram Language for their own customers.

There's an expense setting this up, but a payout in reducing the day-to-day customer support and customer satisfaction.

I bet there are staffers who could implement this as a one-liner. ;)

POSTED BY: Phil Earnhardt

@Phil Earnhardt, I really like this idea! We have, in fact, had brainstorming sessions about building a Wolfram U portal that would help to archive links to past sessions and courses, certifications, and other useful information. It's good to be reminded there is interest out there. Thanks!

POSTED BY: Jamie Peterson
Posted 11 months ago

Well, good. The really interesting thing for me is having a cloud app that Wolfram U users would regularly be using -- and having some sense of what they could [cloud] deploy for themselves.

One of the amazing bits of code that Wolfram U has is the grading engine for the Elementary Introduction to the Wolfram Language course. Anyone who goes through that course without being amazed what the grading engine is doing is just not conscious! Like many, It's tricky: you must have answers that are being produced computationally, but the implementors could not have anticipated everything users could possibly do. I found myself tweaking my answers -- just to see what the engine would judge as "correct" or "incorrect". IMHO, the only shortcoming with the EIWL grading engine is that the source code is not available.

If/when you do this project, please design and implement it with the intention of making the Wolfram Language Notebook source publicly available. Students regularly using it should be curious about its implementation. And people will want to borrow/modify it for their own use; that's what you want. You may even want to do a BigMarker presentation where you walk through the source and explain what/why you do things (both for functionality and for code efficiency).

POSTED BY: Phil Earnhardt
Posted 11 months ago

Trying to understand what Trace is telling me here. I have your words but still do not quite follow. Is it possible to step through this?

Attachment

Attachments:
POSTED BY: Coe Miles

Sure—down the list:

  • The input (with explicit parentheses on the composite function)
  • The Map having been applied takes the composition function and "distributes" it to each term in the nested list, at top level (just f/@{a,b} === {f[a],f[b]})
  • The composite operation/function (N@*Mean) is acting on the first list {1,2,4}, which means that it should be applied like N[Mean[{1,2,4}]]. This is because a composite function f@*g—"the pure operator form of taking g of a thing and then taking f of the result"—acts such that f@*g on a means f[g[a]]. This is the definition of function composition, generally speaking. It's just that in Wolfram Language, we don't write f ∘ g, but f@*g.
  • The expression N[Mean[{1,2,4}]] is evaluated in standard order, from the inside out—first, the Mean is taken, giving 7/3, then N of that result is taken, giving 2.333...
  • The same series of operations is performed on the second list
  • The two results are provided with a List head/wrapper, as that's how Map works (cf. f/@(a+b) === f[a]+f[b] === Plus[f[a],f[b]])

Hopefully that helps, but let me know if not.

POSTED BY: Arben Kalziqi

Folks, it was indeed a trivial yet embarrassing issue—I still had expr defined as a + (b c)/d from last week's lecture :|. You can just do this, and it'll work:

integralAPI = APIFunction[
   {"func" -> "MathExpression"},
   "The integral of " <> ToString[#func] <> " is " <> 
     ToString@Integrate[#func, x] &
   ];

integralAPIDeployed = CloudPublish[
  integralAPI,
  "IntegralAPI"
  ]

As you can see in my deployed version here: https://www.wolframcloud.com/obj/online-courses/PracticalProgramming/IntegralAPI?func=sinx

POSTED BY: Arben Kalziqi

Hi Arben;

Last week you showed an example of Function[] used with a variable as well as a #. Was it your intention to compare a prime function and a non-prime function or a prime function using a # compared to another prime function using a name?

Thanks,

Mitch Sandlin

POSTED BY: Mitchell Sandlin

Hey Mitch—"pure" function! But the idea was primarily just to show the "spectrum" from a full, named function with named variables down to a function with no name and whose variables were unnamed. Function of any kind is technically a pure/anonymous function, even if you provide "dummy variable" names for the inputs just like you would in a definition like myNameFunc[x_]:=....

The progression I have in mind is:

mySineAdder[num_]:=num+Sin[num]
Function[num,num+Sin[num]]
Function[#+Sin[#]]
#+Sin[#]&

The first function is a named ("nonymous", you could say, had history taken a different turn). The last three are pure or "anonymous" functions.

POSTED BY: Arben Kalziqi
Posted 11 months ago

Hi Arben and all, I have a question about the last lesson about complexity time on function definition. I have done this notebook and I want to know if there is a simpler complexity way to code the problem.

PLEASE SEE THE NOTEBOOK ATTACHED TO THIS COMMENT

Thanks
Mauro

Attachments:
POSTED BY: Mauro Bertani

Hey Mauro—while I could probably figure this out if I sat and looked at it for a bit of time, I think I need to set the precedent that I can't take arbitrary code and rewrite it for you. However, what I am happy to do is to take an explanation of what your code is actually doing, with some examples, and tell you how I'd write a function that performs the operation in question.

POSTED BY: Arben Kalziqi
Posted 11 months ago

Ok, Arben. I think to have understood where I have made a mistake. It's very fuzzy, how speak with someone change our prospective, by only to expose the problem. Thanks Mauro

POSTED BY: Mauro Bertani

I understand—trying to figure out how to frame and phrase the problem to somebody else is often half the battle!

POSTED BY: Arben Kalziqi

Hey Arben,
In the Friday lecture you did a quick and dirty example of error trapping:

polygonMaker[n_Integer ] := Graphics[{Red, RegularPolygon[n]}]

Then you added a new statement in real time in your Notebook

polygonMaker[___] := "please enter a digit"

I thought basically the three dashes stand for "anything". But it seems in this context they stand for "anything else".
I don't understand the precedence here. How did "else" get implied? I would have thought it redefined the function polygonMaker to always say "please enter a digit" from now on, regardless of the argument. What happened here?

POSTED BY: Carl Hahn

I believe this is how it works: polygonMaker had two definitions entered into the database. When the function is called, the first version entered is called first, but if the type of the input type isn't an integer, then the next definition is called, which does have a response to anything else. The point is that the domain of polygonMaker was extended by adding that second step, but that extension is only checked when needed.(Technically it would be better for polygonMaker to catch integers less than three)

Here's another example to ponder (entered in successive cells).

f[x_Integer] := x*1.0  
f[x_Real] := IntegerPart[x]  
f[3]  
Out[187]= 3. 
f[3.0]. (* the first version can't be applied to 3.0 so the second version is called *)  
Out[188]= 3
POSTED BY: David Snyder

This is basically it—there are two definitions here, and if one doesn't fit it will try the other. Still, Carl's question of precedence is valid, because ___ really means anything given that it's "zero or more things". The answer is: precedence is given in order from most to least specific definitions.

POSTED BY: Arben Kalziqi

I had not picked up on that nuance even though I have used it (subconsciously?) as in the examples on recursion. I had not formally articulated in my head. WL does not pay attention to the sequence of definitions in code first, but yes to the hierarchy of precedence, which of course it does. Particular takes precedence over global. It's a grammar thing.

But if definitions have equal precedence, then it picks the most recent one. (Correct?)

POSTED BY: Carl Hahn

The documentation page linked by Rohit has this to say:

In the factorial function example used above, it is clear which rule is more general. Often, however, there is no definite ordering in generality of the rules you give. In such cases, the Wolfram System simply tries the rules in the order you give them.

You can check the order that definitions are stored—and tried—by evaluating ?func. (The fact that I had to check the documentation to be sure about this case should tell you something about its relative frequency, I hope!)

POSTED BY: Arben Kalziqi
Posted 11 months ago

For details on precedence / ordering see the documentation.

POSTED BY: Rohit Namjoshi
Posted 11 months ago

I wonder why the /@ stayed pupleRed in the last example of "Good Coding practices" (12. Pure Functions). See attached screenshot

Attachment

Attachments:
POSTED BY: B. Cornas

What I think happened is that he didn't put a space after the "/@" as he was doing it on the fly. It still ran on the fly. If you type it in yourself and put the space in there, it won't be red anymore. You might also see what happens to the # if you deleted the following &, it will turn red due to the syntax error.

POSTED BY: David Snyder
Posted 11 months ago

Thanks :-)

POSTED BY: B. Cornas

The spaces are really just a visual choice made by some programmers—I tend not to use them, and it shouldn't have this effect. I'll try to remember to check the full cell expression after today's class and see if I can figure out what's going on—if you look at the recording, you may note that I didn't expect that to look like that and am messing with it a little to investigate!

EDIT: Yeah, it's just written into the cell manually that that should have a particular background color. If you start typing into that space/editing it, you'll retain the color choice.

POSTED BY: Arben Kalziqi
Posted 11 months ago

Hi Arben, I do not understand what you mean by your last lines -after the EDIT:

I want to use different text colors for the comments in one cell. But of course still bracketet by (.........) to indicate that it is a comment. As some comments are temporarely (like (* FIX this later *) and other comments are permament, I would like different colors for the different kind of comments. As everything is adjustable in Mathematica, I would be surprised if this is not possible. Thanks, Bert

POSTED BY: B. Cornas

In that particular context, I meant that if you press (on mac) cmd + shift + e to see the full cell expression, you could see that the color was "baked in" there. I'd generally recommend only using this for clever little modifications, not as a "primary" means to do things (generally). But it can often provide some insight. As for coloring comments, I recommend just using the Style menu. (Maybe we already dismissed this for some reason, but if so I don't remember!)

POSTED BY: Arben Kalziqi

What I discovered on the MAC is that if you highlight just the text in the comment, not the (* *) at the beginning and the end of the comment, you can then go to the FORMAT drop down menu and select "Text Color". That will change only the color of the comment you are highlighting.

POSTED BY: Carl Hahn
Posted 11 months ago

I don't see an option to start a new Topic, so I just reply. I just saw Ahmed Elbanna's post, 4 days ago, and I agree :-)

Great webinars series. Arben is the absolute BEST presenter I have seen in all the webinars I have already watched. He does it just right on every level and in every detail. Arben, you're a fantastic teacher. You opened my eyes & understanding to new levels in things I already knew. I hope you'll be doing much more presentations.

My question : Is there a way to change the default color of a comment (* Comment *)? I can change the Size, Face, but not the color. This would be really convienient as to make an easy distinction between comments : * Explaining the code and function * Things that need repair, seriously * Things that need brushing up or speed enhancement * Alternative versions * Spots in the code that need extensions in the future * Etc.

It would really be great to have this color function. I know that I could do this with 'text' parts, but that will break the possibility to run the complete code with one Shift+ Enter.

POSTED BY: B. Cornas
Posted 11 months ago

Re. changing the default color of comments look at Preferences / Appearance

enter image description here

POSTED BY: Rohit Namjoshi
Posted 11 months ago

Hi Rohit, I just saw your answer to my first question a minute ago, while going through the list. I got a bit lost in the beginning. I actually now remember having found that it is possible to change the color for the comments. But what I am actually looking for is on the fly use different colors for different kind of comments. E,g. color 1 for code annotation Color 2 for alternatives Color 3 for comments like "Can be done better and faster" Etc.

POSTED BY: B. Cornas

Great course! (As usual from Wolfram). I really like the overall structure, the examples, the smaller checkpoints and the exercises at the end of each day.

Here are some improvements for the 5th exercise of Day 4

txt = ExampleData[{"Text", "DeclarationOfIndependence"}]
    (LetterCounts[txt, 2] // ReverseSort)[[;; 20]]

The solution provided uses Take[.., 20], this takes the first 20 elements. Not the 20 most frequent ones. Adding ReverseSort post fix style, sorts the association biggest->smallest. Similar with the second part.

((WordCounts[txt // DeleteStopwords] // ReverseSort))[[;; 20]]
POSTED BY: Fredrik Doberl

Let me know if I have this wrong, but I believe that LetterCounts is always automatically reverse-sorted by value and thus the first n elements—as taken by Take, for example—should always be the 20 most frequent elements. Do you have a counterexample in mind? (Or perhaps this behavior was different on an older version, perhaps pre-13 or even pre-12?)

EDIT: Oh, I see—there's an issue with the version where stopwords are deleted. Thanks for pointing that out; I'll fix it!

POSTED BY: Arben Kalziqi

Arben,

First thank you for teaching this class. This has been very helpful.

I am running into some unexpected behavior. I just downloaded Version 14, and when I do TreeForm of a list I get:

enter image description here

but when I do the same thing using Plus, I get:

enter image description here

I was expecting to get the same thing but with List replaced with Plus. I am a bit puzzled.

Many thanks for your help. Michael O'Connor

POSTED BY: Michael O'Connor

Hey Michael! The reason this happens is the order of evaluation. I have two suggestions for you to try which I think will clarify this behavior:

  1. Trace[TreeForm[Plus[1,2]]]

  2. TreeForm[List[a,b]] (=== TreeForm[{a,b}])

Let me know if this illuminates the difference for you.

POSTED BY: Arben Kalziqi

Also, Plus has the NumericFunction attribute whereas List does not. Thus one sees the following behavior from TreeForm:

enter image description here

POSTED BY: David Snyder

Arben, this does help! Thank you very much!!!

POSTED BY: Michael O'Connor
Posted 11 months ago

From yesterday's lesson, does anyone understand why

    larger[x_, y_] := If[x > y, x, y];
    Fold[larger, {3, 1, 2, 6, 5}]

returns larger[larger[larger[larger[3, 1], 2], 6], 5] and not 6?

POSTED BY: Coe Miles
Posted 11 months ago

I get 6 on my notebook.

POSTED BY: Steven Brawer

As do I. Coe, you might have given larger an incorrect definition that hasn't been overwritten for one reason or another—try restarting your kernel and running these two lines again? (Re: the new comment—the semicolon is fine. It's not necessary for SetDelayed definitions since the evaluation is, as claimed, delayed!)

POSTED BY: Arben Kalziqi
Posted 11 months ago

Try removing the semicolon

POSTED BY: Steven Brawer
Posted 11 months ago

map mystery

Can someone explain what is going on in the expressions below the tree, or would this be too advanced for the level of this course (which, if it is, is OK with me)?

(See attached png file)

Attachment

Attachments:
POSTED BY: Steven Brawer

Hey Steven—I think that these three points in concert are enough to answer your question, but please let me know if they don't:

  • Map[f,expr,n] maps f to expr down to level n
  • Map[f,expr,{n}] maps f to expr only at level n
  • Map[f,x] = x for AtomQ[x]==True (cf. our discussion yesterday: Map goes into the inside of an expression to work on its parts; atoms don't have parts so Map has nowhere to put the f—this could have been designed to produce an error message, but the decision was to have Map function such that if it can't map a function f somewhere, it just returns the object onto which f was asked to be mapped)
POSTED BY: Arben Kalziqi

Hey Arben,

Loving your class. Got a little exercise:

The transfer function of non linear power amplifiers produces inter-modulation products between the tones at their inputs. If the transfer function is expressed as a simple 5th order odd polynomial, and the input x is made of two sine waves at frequencies w1 and w2, you can easily find the mess that shows up at the output. It includes 3rd and 5th order inter-modulation products around the two tones, 3rd and 5th harmonics, and inter-modulation products around the harmonics. In RF systems we use band-pass filters to select only the fundamental tones and those tones around the fundamentals w1 and w2 (usually assumed to be closely spaced compared to the frequencies of the harmonics... Although audio systems will not be so cooperative.)

I've been playing around with some of the concepts I thought I was learning to take the output and literally select them, but it's not working yet. In the past I've just given up and brute forced it (yuk). This time round I'd like to learn how to do it right...help?

POSTED BY: Carl Hahn

We really do abhor a brute force solution. This is a fun little question; please see embedded response!

PLEASE SEE THE NOTEBOOK ATTACHED TO THIS COMMENT

Attachments:
POSTED BY: Arben Kalziqi

Thank you, Arben!!!

That's a really neat approach. Had not yet thought along the path of creating a list of all the terms. But entirely consistent with thinking about the heads and changing them. That's a powerful analytical tool.

Your solution has a lot of pedagogical value. I am going to be staring at it a bit.

If you are on a roll, I've got another one coming as soon as I can figure out how to ask it properly. Let me take a crack and if it needs more let me know: It has to do with a problem in FM theory, where recognizing that Cos (B Sin [w t]) can be expanded as a Fourier series whose coefficients are an ordinary Bessel Function of the first kind

= Jo(B) + 2 Sum (J2n(B) Cos (2 n w t) )

Then you have to truncate the series to get the coefficients of interest. There is a similar expansion for Sin(B Sin (w t). You can sample an FM signal and do an FFT and just see the spectrum, but recognizing the Bessel Function behavior provides an intuitive understanding of FM theory (bandwidth, mod index, S/N properties etc) that even when FM was really popular few people grasped. Now it's beginning to be a lost art. What I never figured out how to do was use WL pattern matching and mapping tools to "recognize" and derive this classical solution. It has other analytical advantages for other problems too.

POSTED BY: Carl Hahn

I really strive to provide pedagogically useful answers, so I appreciate it :). If I'm reading your (original) post correctly, you could also do pattern-matching or selection by going directly into the cosine terms, extracting the coefficients of the omegas with CoefficientList and grabbing only those cosine terms for which the absolute value of the difference is 1. This would make it so you don't even have to generate the desiredHarmonics list, if that's indeed what you want.

As for your new question, some more information would indeed be helpful. It's been a long time since I've dealt with this sort of thing, and certainly I've never approached it from the perspective of signal analysis and decomposition. But, fundamentally, you could write the same kind of thing as above with Cases, specifically recalling that Cases allows not just for pattern-matching but for term-rewriting too. So you could do something like:

Cases[expr,Cos[B___ Sin[ω t]]:>(full or truncated B-dependent sum here)]

This would extract all of the relevant cosine terms and rewrite them; you could update the rule to include both sine and cosine terms but I'm forgetting the way to do that right now. You could do something similar to the previous case and just extract all cosine or sine terms that have the right form, then use a simple /. to write a rule which appropriately converts the cosines and sines to their sum forms. You may also be interested in TruncateSum (new in 14), though this behavior was always manually implementable with a little elbow grease.

Otherwise, if you have an exact worked example, I'm pretty sure I could write something up reasonably quickly.

POSTED BY: Arben Kalziqi

Thanks Arben, I will get back to you on that FM question. I have to find my notes on when I last tried to do it.

"If I'm reading your (original) post correctly, you could also do pattern-matching or selection by going directly into the cosine terms, extracting the coefficients of the omegas with CoefficientList and grabbing only those cosine terms for which the absolute value of the difference is 1"

That sounds kinda like what I was originally trying to do but unsuccessfully. Could you show an example?

POSTED BY: Carl Hahn

Let's see:

PLEASE SEE THE NOTEBOOK ATTACHED TO THIS COMMENT

EDIT: Really better just to use Coefficient. I don't like CoefficientList anymore; it's dead to me. So this is probably the most efficiently and extensibly you could write it?

Cases[yexp, 
 c___*Cos[arg_] /; 
  Abs@Total@Coefficient[arg, {ω1, ω2}] == 1]
Attachments:
POSTED BY: Arben Kalziqi

Delicious!!! Once again a pedagogical Rosetta Stone. Thank you , Carl

POSTED BY: Carl Hahn
Posted 11 months ago

Hey everyone, care to take a stab at explaining why:

Total[{{a, b, c}, {u, v, w}, {x, y, z}}]

returns

{a + u + x, b + v + y, c + w + z}  (*One of our Day 4 "Check Your Understanding" questions*)

I've looked at the documentation for "Total" and still can't figure out how this result is obtained. Thanks in advance,

POSTED BY: Coe Miles

Hey Coe, I can explain this.

Total is documented to take the elements of a list and add them together. In other words:

Total[{a,b,c,...}] === Plus@@{a,b,c,...} === Plus[a,b,c,...]

In standard notation, that means that Total[{a,b,c,...}] just evaluates to a+b+c+.... Nothing too surprising here, I think.

The misunderstanding likely comes from two places.

Since Wolfram Language is fully symbolic by default, we never said what a, b, c and so on might be. They don't need to be reals or integers or even numbers—in fact, try this for me:

randomImages=Table[RandomImage[{0, 1}, {25, 25}], {2}]
Total@randomImages

If there is a sense in which the arguments inside the List fed to Total can be added, then Total will add them because it is very literally just performing the pattern-matching term-rewriting transformation rule I expressed above.

Now that we know what Total does, if we want to understand the output of Total[{{a, b, c}, {u, v, w}, {x, y, z}}], we need to ask: what does Plus do when fed multiple List objects?

You can try this out manually, but that's basically just rewriting the thing that we're not understanding in the first place. Instead, you might try Attributes[Plus]. You get back several attributes, but the important one here is Listable. If you check the documentation for that, you'll see examples which basically replicate the behavior that prompted this question:

In[1]:= {a, b, c} + {x, y, z}

Out[1]= {a + x, b + y, c + z}

This is to say: Listable functions automatically thread over their arguments when those arguments are lists. Most built-in mathematics functions are listable, and this attribute allows for highly optimized code compared to looping or even mapping.

In the case of Plus, this behavior also happens to correspond to the standard mathematical operation of addition acting on vectors, matrices, and tensors more generally. After all, if you have the vectors a, b, and c, you find their sum by element-wise addition, right? The same is true for matrices and so on. (And in fact, people are often lazy and will "add" a scalar to a vector when what they really mean is that they want to add a constant vector to a vector. For example, somebody might say x = <1,2,3>, then say something like x + 1, when really they mean x + <1,1,1>. This is also automatically handled by Listable; feel free to evaluate {a,b,c}+1 and see the result, or even {{a,b},{c,d}}+1.)

POSTED BY: Arben Kalziqi
Posted 11 months ago

Arben - Thank you very much for the detailed explanation. I am finding this material somewhat difficult but rich and rewarding when something sticks.

POSTED BY: Coe Miles
Posted 11 months ago

Tree walking - inconsistency or design

Given some algebraic expression y, I can do y[[3,2,5,1,..]] to walk the tree (or the full form) to a particular node.I can also do y[[0]] which is the same as y//Head. However, y[[0,2]], for example, is an error in my notebook, though presumably it should be the same as y[[2]].

This seems to me an inconsistency, but perhaps it is that way on purpose???? (In y[[3,2,5,1,..]], presumably the "0" is implicit.)

This is probably not a big deal, but I am bothered by stuff like this - a sort-of reflection edge condition that doesn't seem sensible.

POSTED BY: Steven Brawer

I don't think that this is an inconsistency: if you do extract a head with [[0]], that head does not have some internal structure that you can probe by asking to get its parts. What is the second part of Graphics?

POSTED BY: Arben Kalziqi
Posted 11 months ago

It seems to me the Head is just a node. I don't understand why it should be different from any other node. If it doesn't have a structure, why should any other node have a structure? I suppose there is something in the engine which interprets the head node differently from other nodes.

POSTED BY: Steven Brawer

I think I see what you mean; I think that that's fair enough if you're thinking about it in terms of TreeForm. The discrepancy, such as it is, exists because technically Part is always "reaching into" an expression (cf. the reason you can't [""] use Part on atoms is because they don't have an "inside" to reach into).

That is: whenever you say expr[[n]], you are going into expr and extracting something. If you have not just a single n, but exprp[[m,n,...]] then the number of elements in that sequence specifies the Depth at which you are accessing the structure of expr. The issue here is that expr[[0]] does indeed not work like this: because indexing in Wolfram Language starts at 1, this is not "getting a part" of expr in the same sense that expr[[n]] for n≥1 is, but is "special case", manually-implemented extra behavior.

I do see why one might consider this inconsistent, but my personal sense is that [[0]] is simply used as a convenient syntax shortcut for Head@. If we didn't have it, the behavior of Part would indeed be more consistent, but it would also be less functional for some particular uses... my perhaps-frustrating advice is to do a bit of compartmentalization about the meaning of Part, here. We're already doing that when using Part on SparseArray or Dataset atoms, so why not add one more case where Part is doing something which is maybe a little bit misleading...

Finally, I do want to clarify something: because this [[0]] is a special behavior which does not give you the 0th part in the technical sense (as there is no 0th part in a language which indexes from 1), it is not the case that there is an implicit 0 which starts off any Part sequence. Moreover, there's nothing in any given expr that you cannot access because of this: the entirety of the internal structure of expr (if it has any) is accessible without making any reference to any 0th part. It's just that if you would like to get not the internal structure, but the thing that wraps around that internal structure—which in the strict sense is not a part—then we allow you to do that with Part[expr,0] as a convenience.

The succinct version is maybe:

  • Part is really only intended to talk about the parts of a compound expression, and the "parts" of an expression are always internal to the expression.
  • Because you could already access everything about an expression but its head with Part, we added a special case of asking for """the zeroth part""" so that you could get that last bit of information without having to leave Part-world.
POSTED BY: Arben Kalziqi
Posted 11 months ago

Thanks. That's really helpful.

POSTED BY: Steven Brawer
Posted 11 months ago

In the "Check Your Understanding" exercise for pure functions we are told that Function[u, 3+#][x] is the correct syntax for returning 3+x given x as an argument.

I think that's incorrect (?)

POSTED BY: Paul Tikotin

Good catch! I've fixed this in our repo and it will be fixed in our next deployment.

POSTED BY: Arben Kalziqi
Posted 11 months ago

I asked in the BigMarker Q&A panel how to upload a WL Notebook to my Wolfram Cloud account. The reply was to use the "upload" button in the upper right corner of the Wolfram Cloud window. This doesn't work; I get the error message, "To access this feature, subscribe to a plan." Apparently, the Basic plan isn't good enough.

I can upload files through Mathematica on another machine through the File->Save to Cloud function. I don't understand why that works but simply doing an "upload" in cloud does not. Is there a good reason for keeping a random user from uploading any notebook to a Wolfram Cloud account through the web? If not, please file this as a bug.

One note for the staff answering questions: when they are answering a question like this, they should try to perform the function with the same level of privilege/access as a normal customer. In this case, the Wolfram Cloud "upload" function should have been tried on a Basic plan.

POSTED BY: Phil Earnhardt

Hi Phil,

I think this can be better understood and handled through contacting customer support:
https://www.wolfram.com/support/contact/email

POSTED BY: Ahmed Elbanna

Hello Phil,

Thank you for letting me know. I apologize for the confusion created.

Upon a detailed look, the behavior described seems to be related to the fact that Wolfram Cloud and Wolfram Desktop are two separate products. The basic plan for Cloud grants users the basic functionalities needed for a smooth Desktop operation, but it does not grant access to the Wolfram Cloud as a product in and of itself. Basic Plan restrictions

This is why the direct dynamic interaction within Wolfram Cloud is not available but the use of CloudPublish from Desktop, either programmatically or through the menu tool, works as expected.

I will try to improve my Q&A support with your feedback. I have already setup a basic account for testing of future questions. Thank you!

POSTED BY: Oliver Jandette
Posted 11 months ago

Hi Everybody!

I would like to produce documents effectively (so programmatically) and I need to find help about producing cells, sometimes cells within cells (within cells?). Can anybody give a good reference? Such things as TextData, BoxData, ToBoxes, ToString and similar.

POSTED BY: Artur Piekosz

Yes, I also found that I couldn't use the code output by those boxes in situ. Entering it into a new code cell works fine.

POSTED BY: Arben Kalziqi

To produce cells look at Cell, TextCell and ExpressionCell. Here's my little experiment on WolframCloud to programmatically produce Hierarchical Textcells with Section and Subsection styling and CellGroup function.

enter image description here

and the result is this

enter image description here

Cheers.

POSTED BY: Jack I Houng

I'd highly recommend Professor Richard J. Gaylord's Wolfram Language Fundamentals

Posted 11 months ago

Hate to redundant, I still can't get by this. If this is a real error, other errors I get may be me or Mathematica - can't tell. Btw, $Version returns: 14.0.0 for Mac OS X ARM (64-bit) (December 13, 2023).
enter image description here

POSTED BY: Coe Miles

Please see the answers downthread—the easiest way I know to explain this is that if you say ax, how is Wolfram Language to know that you actually, truly, really meant a*x? It can't.

If you were to type myVar^2, you ostensibly would not expect to get m^2 * y^2 * V^2 * a^2 * r^2, so why expect differently for ax^2?

(Since tone is never conveyed well over text, I want to be clear that the intended tone here is "Socratic" rather than "snippy" ;).)

POSTED BY: Arben Kalziqi
Posted 11 months ago

ax is on variable not two variables a * x is and expression equal to a space x

POSTED BY: Gregory Sowder
Posted 11 months ago

I think the best way to see what is going on, on your own machine, is to look at y//FullForm and z//FullForm. They are quite different.

enter image description here

However if you insert a space between a and x in your expression for z, then y and z should be the same.

[If you are somehow stripping out spaces, then that will cause you problems. Mathematica really uses the spaces as Arben says.

I am already finding that when I am getting odd results, using FullForm, TreeForm etc can make things clearer.

POSTED BY: Paul Tikotin
Posted 11 months ago

Perfect. I get it, thank you.

POSTED BY: Coe Miles

While working with the exercises for Day 3 (Syntax and Expressions), I got the following error message while attempting Exercise 4 (ListLinePlot for airline passenger data), namely "An improperly formatted option was encountered. The left-hand side of the option was not a symbol or string* My solution(s) worked fine initially, but then went wonky all of sudden. In fact every plot in the notebook threw the same error message, whether options were specified or not. (Also all of the dynamic Hints and Solutions boxes disappeared, but that's not my main concern right now). Please see attached notebook..

Attachments:
POSTED BY: Michael Ulrey

Hi Michael—I think the issue comes from the formatting that this notebook has for that ListLinePlot input. Ideally, it would just be an input... but when you click into it, you can see that there's a kind of yellow box around it. Basically, this is a text cell formatted as input and so it won't evaluate correctly, if we're seeing the same thing. (I think your specific errors with respect to Dynamic elements here are kernel-specific and thus not "carried with" the notebook that you uploaded when I open it on my end, so I can't see for sure what your issue was.)

if you just start typing ListLinePlot in a new cell, you should be able to plot the data without issue. I think what's happening is that on our end, making the Hint/Solution buttons does something unfortunate to the formatting of code blocks that we had intended for you to type into. I'll look into this with our team.

POSTED BY: Arben Kalziqi

Arben, Thanks, that makes a lot of sense since both errors occurred at the same time. I'll try out your suggestions and see what happens.

POSTED BY: Michael Ulrey
Posted 11 months ago

Yes, I also found that I couldn't use the code output by those boxes in situ. Entering it into a new code cell works fine.

POSTED BY: Paul Tikotin

Arben, Just to let you know I opened this notebook the next day in order to try your suggestions, and it had healed itself overnight. One of those situations you're glad it's fixed without any effort, but not quite sure why. Anyway, all is good for now, and everything is working again as expected. I even entered some new comments and code into the notebook, and nothing seemed to upset it again. Fingers crossed! :<) :<)

POSTED BY: Michael Ulrey
Posted 11 months ago
In[84]:= Clear[x, y, a, b, c]

In[85]:= y = a x^2 + b x + c

Out[85]= c + b x + a x^2

In[86]:= Expand[y^2]

Out[86]= c^2 + 2 b c x + b^2 x^2 + 2 a c x^2 + 2 a b x^3 + a^2 x^4

system: Apple M2 Ultra, 14.3 (23D56) Mathematica: 14.0.0.0 Mac OS X ARM (64-bit)

POSTED BY: Randy Janke

Is anyone else experiencing problems with the question submission bar in bigmarker on Chrome? When I go to type, the text box turns white, with white text, rendering the text unreadable.

Posted 11 months ago

I know you touched on this, but in exp = 1 + x^2, exp[[2, 1]] returns x, yet exp[[1, 1]] is longer than the depth of the object, why again?

POSTED BY: Updating Name

exp[[2, 2]] returns 2, the second part of the second part, I'd expect exp[[1, 1]] to return the first and only part of the first.

In Wolfram Language, the structure of expressions and how they are indexed can sometimes be counterintuitive if you're not familiar with the underlying representation. Let's break down your example to understand it better.

Consider the expression exp = 1 + x^2. In Wolfram Language, this is internally represented as Plus[1, Power[x, 2]]. The head of this expression is Plus, and it has two arguments: 1 and Power[x, 2].

exp[[2, 1]] returns x because it's accessing the first part (x) of the second argument (Power[x, 2]) of exp. exp[[2, 2]] returns 2 because it's accessing the second part (2) of the second argument (Power[x, 2]) of exp. Now, regarding exp[[1, 1]], the confusion arises because 1 is an atomic expression in Wolfram Language (it does not have a head and arguments structure like Power[x, 2] does). Atomic expressions include things like numbers, symbols, and strings. When you try to access a part of an atomic expression, it results in an error or an unexpected result because atomic expressions are considered as having no parts.

So, exp[[1, 1]] is trying to access a part of the atomic expression 1, which doesn't have any parts, hence the issue. In Wolfram Language, the parts of atomic expressions are not accessible in the same way as the parts of composite expressions.

  • GPT 4

[redacted]

AtomQ[#] & /@ {exp[[1]], exp[[2]]} returns {True, False}, and Atoms don't have parts (errrrrrr... it's counterintuitive).

Posted 11 months ago

...and this can be seen: AtomQ[exp[[1]]] returns true.

POSTED BY: Paul Tikotin

Yeah, it's just atoms have parts as far as atoms are concerned, as apposed to atomic expressions. The same signifier ("atom") points to different signified concepts depending on the context, hence counterintuitive. Essentially, atomic parts are not the same in the case of the fundamental concept in chemistry and physics and atomic expressions and therefore not covered by Part in the normative sense. So an error message is returned.

There are sub-atomic parts

Understanding this helps in navigating subjects that cross disciplinary boundaries. - GPT4

Hi Lewis—I understand what you're saying abstractly, but am having trouble understanding how it applies into specific to the discussion of atoms today. Let me offer a few clarifying comments and we can go from there?

  • Atoms are the fundamental building blocks of compound expressions in Wolfram Language.
  • Any compound expression, when you get to "the bottom of it"—literally, in the sense of TreeForm—is made up of atoms.
  • "Atom", in the context of Wolfram Language, is meant in the original Greek sense: "a thing which cannot be cut". What we mean by "uncuttable" is that Part/[[ ]] does not fundamentally work on atoms; various functions can pull things out of atomic objects (cf. Numerator and Denominator for Rational atoms, ReIm for Complex atoms, EdgeList for Graph atoms, and so on).
  • However—and it's a big "however"—there are some cases where an atomic object might be expected to have its insides accessible via Part (e.g. Dataset, SparseArray, NumericArray...). To that end, the front end will automatically "translate" a user request for Part into the correct underlying code to give the "expected" result. This lets you access the parts of a SparseArray "with" Part, for example.
  • Ultimately, though, this is a syntactic convenience—just like, in a sense, being able to say a+b rather than Plus[a,b] or {a,b,c} rather than List[a,b,c]. There exists an underlying "truth" and Part is not part of that truth when it comes to atoms.
POSTED BY: Arben Kalziqi
Posted 11 months ago

Nice! I hadn't thought (until today!) about the derivation of the word atom. So, 'tom' as in 'tomography'.

POSTED BY: Paul Tikotin

Yes, the very same! It's from τόμος, which I think is from τέμνω, which is something like "to cut/slice/section/portion".

POSTED BY: Arben Kalziqi

The notion of uncuttable is really helpful, thank you. I’ve now learnt to turn to inbuilt functions in order to better understand what’s going on and just the idea of syntactic convenience is helpful in getting over what at times has felt like contradictions in terms. I remember Prof Richard J. Gaylord saying once that Wolfram, at times, behaves strange, and never quite knew what was meant by that (likely the pitfalls of intuition).

Posted 11 months ago

I think the apparent confusion regarding the word atom is as a result of an error in the evolution of our scientific knowledge. It is only relatively recently that we discovered that 'atoms' of, say, radium are not really atoms in the logical sense of the word. Then we repeated the mistake, calling some of the sub-atomic bits "fundamental" only to find...

In a parallel universe where the Wolfram language was developed in 1880, the concepts of atoms in Mathematica and Chemistry would have been consistent. Although some physicists maintained that the atoms of the chemical elements were fictional conveniences for the sake of calculation and that, if they even existed they would be too small to ever be known. Then came 1905 and Einstein's analysis of Brownian motion.

POSTED BY: Paul Tikotin

One thing that really attracted me to Wolfram was how the expressions in a book I picked up published in 1988 still computed. Thankful, in this universe there is forward compatibility, even amongst changes in scientific knowledge.

I'm with you there—I immediately thought that the following image, from this blog post, was very striking:

enter image description here

POSTED BY: Arben Kalziqi

Although Show[Graphics3D[Stellate[ Icosahedron[ ]]]] wasn't forward compatible.

Heh, it's true :). At least in this case it's been superseded by something simpler and clearer!

POSTED BY: Arben Kalziqi
Posted 11 months ago

I can heartily recommend this. While I am in no position to judge Wolfram's contribution to Physics, the ideas are fascinating and are something I would like to pursue.

If I am hearing him correctly, he is attempting to make sense of Physics by looking at it as a system that is computable by the application of rules for re-writing. He seems to be saying the system evolves what look like small black holes and "elementary" particles from vacuum energy. Of course he is talking to a professional physicist here so we hear him respond to some good questions.

I am stricken by the similarity of the method Wolfram is proposing and the spirit of the the approach we seem to be taking in grappling with Mathematica. Very stimulating!

https://www.youtube.com/live/ITJ3AF3TK5M?si=rECCBuBAAVWa_Im4

POSTED BY: Paul Tikotin
Posted 11 months ago

Steven/Arben - I too am getting the INCORRECT result. I am also running Mathematica 14.0.0 on Mac silicon. (This might also explain some of the results I have been getting!) I create a notebook and set y = ax^2 + bx + c. I then try Expand[y^2] and get the following:

    ax^4 + 2 ax^2 bx + bx^2 + 2 ax^2 c + 2 bx c + c^2

I have attached my notebook and an image making it clear what element is incorrect in Mathematica's answer.

Attachment

Attachments:
POSTED BY: Coe Miles
Posted 11 months ago

PROBLEM PROBLEM PROBLEM

When I create a notebook and try y = ax^2 + bx + c, then

Expand[y^2] gives
ax^4 + 2 ax^2 bx + bx^2 + 2 ax^2 c + 2 bx c + c^2
in my notebook, which is incorrect,

while yours has
c^2 + 2 b c x + b^2 x^2 + 2 a c x^2 + 2 a b x^3 + a^2 x^4
which is correct.

I'm, using Mathematica 14.0.0 on Mac silicon. Here is my notebook

POSTED BY: Steven Brawer
Posted 11 months ago

More info: When I do "y = ax^2 + bx + c", with explicit asterisk, then Expand[y^2] is correct c^2 + 2 b c x + b^2 x^2 + 2 a c x^2 + 2 a b x^3 + a^2 x^4

Note: The preview removes the explicit asterisks. What works is y = a asterisk x^2 + b asterisk x + c

POSTED BY: Steven Brawer
Posted 11 months ago

Write "a x" with a space between a and x. Mathematica interprets the space as a multiplication.

POSTED BY: lara wag

Addendum to Lara's explanation: "ax" is read as one symbol, as opposed to "a x" which is read (in the example case) as two distinct symbols of a numeric type, so the space is interpreted by Mathematica as an implied multiplication.

POSTED BY: David Snyder
Posted 11 months ago

I don't believe this is a space issue. The notebook shows input (w/o spaces) and the incorrect output.

POSTED BY: Coe Miles

Hey Coe—I'm not sure what you mean. The attached notebook is missing the spaces between the coefficients and thus produces the incorrect output. You want:

y=a x^2+b x+c

or, explicitly:

y=a*x^2+b*x+c

in order to get the correct polynomial expansion.

@ Steve, you can type your code between two backticks to prevent it from formatting via markdown, which is why the asterisks are being used for formatting rather than displaying.

POSTED BY: Arben Kalziqi
Posted 11 months ago

When I enter everything without spaces I get the wrong answer. When I enter WITH explicit asterisks, I get the right solution. The attached image shows what is entered and the result.

Attachment

Attachments:
POSTED BY: Coe Miles
Posted 11 months ago

Steven - I get the correct answer when using explicit multiplication:

    y = a*x^2 + b*x + c

Why is this? Arben?

POSTED BY: Coe Miles

To be absolutely clear: this is equivalent to asking why var^2 does not return v^2 a^2 r^2. If you don't type a space or an asterisk between two letters, there's no way for the system to know that you're talking about two separate symbols.

POSTED BY: Arben Kalziqi
Posted 11 months ago

Bang head with hand. I forgot. Thanks.

POSTED BY: Steven Brawer

It can happen even after a decade of using the language—I was once thwarted for an entire day by an errant curly/("""smart""") quote breaking a string expression when I first start working here!

POSTED BY: Arben Kalziqi
Posted 11 months ago

Another way of saying the same thing: Mathematica is taking ax^2 to mean (ax)^2. This can be seen by asking for the FullForm.

In terms of today's lesson: AtomQ[ax] returns True , AtomQ[a x] returns False.

POSTED BY: Paul Tikotin
Posted 11 months ago

Hi Arben,

Would please clarify the following terms: vector, array, matrix, expression? And regarding the docs, what's the difference between options and properties in a function?

POSTED BY: Updating Name

When I'm using those terms, I mean:

  • Vector: something that you can describe with 1 index. It could have a length of 2 or a million, but you could get any singular element out of it by doing vec[[n]].
  • Matrix: something that needs two indices to get any individual elements from: say, the element at the first row and third column could be accessed with mat[[1,3]].
  • Tensor: generic term for anything where you'd need more indices to describe the position of a single element. Note that this is not really an appropriate definition in the physics sense, but I'm being a little loose with it. You can think of them as the general case of the special "one-index" and "two-index" cases of "vector" and "matrix", respectively.
  • Expression: any sort of singular "thing" in Wolfram Language—that is, anything that's not a sequence like a,b. a is an expression, Plot[x,{x,0,5}] is an expression, and so on.

You can see more of the fun of conflicting definitions here.

POSTED BY: Arben Kalziqi
Posted 11 months ago

I thought I understood this material during the lecture; most things made sense. Some parts of Problem 1 however just don't seem right. For example, why does pat3 give the following result?

exprs = {1, h[], h[a], h[b], h[a, b], h[a, a], h[a, b, c], g[h[a, b]]}
pat3 = x_
Cases[ exprs, pat3]
    {1, h[], h[a], h[b], h[a, b], h[a, a], h[a, b, c], g[h[a, b]]}

I see no way to get this result.

Thanks in advance,

POSTED BY: Coe Miles

pat3 is x_, which means "any one thing—call it x." Every element of exprs is just a single thing, so they all match. Because we're not using that x anywhere, this pattern is equivalent to just _, i.e. "any one thing".

We might see a named pattern like x_ in a standard function definition:

f[x_]:=x^3

The x on the LHS has nothing to do implicitly with the "any one thing" denoted by the _. The x is just a dummy name that we give to that one thing (the one thing that appears in f's brackets) so that we can refer to it on the RHS. Consider the example:

f[x_]:=5

There's no mention of x on the RHS, so it's pointless to even name it on the LHS (in some sense.) It would be exactly the same to instead say:

f[_]:=5

The point is that "f of any one thing" should be 5. We don't care about the input here, so why bother to name it if we're never going to use it? In this sense, we are saying that f is a transformation rule that takes anything of the form "f of any one input, and I don't care what that input is" and returns 5.

POSTED BY: Arben Kalziqi
Posted 11 months ago

My apologies folks, I mis-wrote the equality that I am asking about. Again -

In today's post-lecture questions "pat2 = _h" Case tells me that pat2 matches the following elements in exprs:

    {h[], h[a], h[b], h[a, b], h[a, a], h[a, b, c]}

Does this mean that **pat2 = _h = h_**? It seems like it, but this seems odd to me in that we are PATTERN MATCHING and, it seems to me, elements in one order should not necessarily be equivalent to elements in a different order.

Comment?

POSTED BY: Coe Miles

pat2 is _h, which explicitly means: "any one thing with head h". This is not the same as h_, which means "any one thing—call it h". It is (except for perhaps some pathological edge cases) the same as h[_], which is "h of any one thing."

POSTED BY: Arben Kalziqi
Posted 11 months ago

In today's post-lecture questions "pat2 = _h" Case tells me that pat2 matches the following elements in exprs:

    {h[], h[a], h[b], h[a, b], h[a, a], h[a, b, c]}

Does this mean that pat2 = h = h? It seems like it, but this seems odd to me in that we are PATTERN MATCHING and, it seems to me, elements in one order should not necessarily be equivalent to elements in a different order.

Comment?

POSTED BY: Coe Miles

question = "If Mathematica is a term rewriting system, when I write a programme in Mathematica, am I writing a term rewriting rule, and is the programme a function?";

If[question==={"yes", "yes"}, "At what point does a rule become a system?","Help!"]

Yes! The evaluation system is effectively a huge database of pattern-matching term-rewriting rules. For example, think about:

D[x^3,x]

We know that this is "the derivative of x cubed with respect to x", and we know that when we see this pattern—the derivative of x^n_, if you will—that we should take the power, drop it to the front, and then subtract 1 from the power.

The Wolfram system is effectively doing the same thing: it sees that something matches the pattern D[x^n_,x] and says what we do: "I know what to do when I see that: D[x_n,x] :> n*x^(n-1) and that's what's returned to us.

When you write your own functions, you are effectively adding to this database of term-rewriting rules (and are often using those rules as foundations for your own definition). In this way, any program you write is just about matching patterns and transforming the appropriate ones into something new.

POSTED BY: Arben Kalziqi

Wonderful explanation! I'm going to give this more thought. I'm trying to get my head around Wolfram engine.

Posted 11 months ago

Anyone grok this list operation: Alphabet[][[;; 10 ;; 4]]

I understand Alphabet[]
I understand Alphabet[] [[;;10]]
I understand Alphabet[] [[;;4]]

I do not grok what's happening when the 2nd and 3rd statements are combined. I thought the ";;" operation acted like a pipe. The solution {a, e, i} isn't what I think a pipe should give.

POSTED BY: Coe Miles

Hey Coe, could you be a little clearer in your description of your inputs, their actual outputs, and the outputs you would expect? In the meantime, let me explain the meaning of each line you've included here.

Alphabet[][[;;10;;4]]

This is equivalent to Alphabet[][[1;;10;;4]]. This means "starting with the list of all of the letters in the (Latin) alphabet, take the first through the tenth elements in steps of 4." I.e., that's elements 1,1+4,1+4,+4===1,5,9. (You don't go to 13 since the initial list only has 10 elements.) And indeed, that's what this code does—cf. MapIndexed[#2[[1]] -> #1 &, Alphabet[]] or Thread[Range@26 -> Alphabet[]] to see it directly.

Actually, I see what you were initially asking. I think the above sufficiently clarifies it, but "in order":

  • Alphabet[] provides a list containing the characters of the Latin alphabet
  • Appending [[;;10]] is the same as appending [[1;;10]] and is extracting the 1st through 10th parts of the list containing the alphabet
  • Similarly, appending [[;;4]] is equivalent to appending [[1;;4]] and is extracting the 1st through 4th parts of the list containing the alphabet

Doing [[a;;b;;c]] is not the same as doing [[a;;b]][[;;c]]. The "sandwich" construction specifically means "go from a to b in jumps of size c"; it's different from "chaining" ;;/Span commands through [[ ]]/Part.

POSTED BY: Arben Kalziqi

Hello Coe!

Arben's explanation made me want to see it in color. This is what came of it. I hope it helps.

PLEASE SEE THE NOTEBOOK ATTACHED TO THIS COMMENT

Have fun!

Attachments:
POSTED BY: Lori Johnson
Posted 11 months ago

Thanks for teaching and I look forward to gaining new insight and learning the various facets of the Wolfram Language and learning to code with Mathematica. The course in this series labeled with #34 for the most part was attended, although completion of the course-work is latent. Hastily, without having differentiated or juxtaposed the two sets of files, I'm wondering about the development or difference thereof. Also, is this the appropriate place to provide commentary or share what you are learning or problems that you may be stuck on in the sense of posing a problem that is confounding?

For the sake of convenience, I have the following two lines for reference with the time zone set to GMT-8 for current locale the default being +1 given the supposed original temporal datum or original time where it is that the Vatican City state is presently, as well as central Europe and Scandinavia. This is just for contrasting disparate times along the lines of DateObject[] and use of SolarTime[] and MoonPosition[] for scheduling, thinking with respect to calendrical arithmetic or geo-computation for arranging one's day according to a plan in the sense of avoiding scheduling conflicts and slotting time around attendance of the course.

Transpose@{Alphabet["Greek"][[1;;10]]//Capitalize,(QuantityMagnitude/@Table[(a-DateObject[List[1,1,11,0,0,0],"Instant","Gregorian",1.`]),{a,{DateObject[{2024,1,29,9,0,0}],DateObject[{2024,1,30,9,0,0}],DateObject[{2024,1,31,9,0,0}],DateObject[{2024,2,1,9,0,0}],DateObject[{2024,2,2,9,0,0}],DateObject[{2024,2,5,9,0,0}],DateObject[{2024,2,6,9,0,0}],DateObject[{2024,2,7,9,0,0}],DateObject[{2024,2,8,9,0,0}],DateObject[{2024,2,9,9,0,0}]}}])//Ceiling//FactorInteger}//Grid

and also

TableForm[Transpose[List[List["Lists and Associations","Assignments, Rules and Patterns","Syntax and Expressions","Writing a Program","Good Coding Practices","Interfaces and Deployment","Working with Data, Part I","Working with Data, Part II", "Developing Packages","Review Session"],List[DateObject[{2024,1,29,9,0,0},"Instant","Gregorian", -8.`],DateObject[{2024,1,30,9,0,0}, "Instant", "Gregorian", -8.`], DateObject[{2024,1,31,9,0,0}, "Instant","Gregorian",-8.`],DateObject[{2024,2,1,9,0,0},"Instant","Gregorian",-8.`],DateObject[{2024,2,2,9,0,0},"Instant","Gregorian",-8.`],DateObject[{2024,2,5,9,0,0},"Instant","Gregorian",-8.`],DateObject[{2024,2,6,9,0,0}, "Instant", "Gregorian", -8.`],DateObject[{2024,2,7,9,0,0}, "Instant", "Gregorian",-8.`],DateObject[{2024,2,8,9,0,0},"Instant","Gregorian",-8.`],DateObject[{2024,2,9,9,0,0}, "Instant","Gregorian", -8.`]]]]]

With respect to this is the question of how to seamlessly integrate sharing of snippets of code like this in and that within Wolfram|One this was drafted and then the sequence was right click the "cell object delineator" (For lack of the proper term, assuming that is the incorrect phrase or string for a potential association) and then through the "drop down menu" "Plain Text" was selected and then "Alt + Tab" to navigate to Notepad and then "Ctrl + V" in the window after left clicking and then copying again via the aforementioned technique and then tabbing or repeatedly using the "Alt + Tab" macro/technique or combined key-press selecting the Microsoft Edge window and then again (As with Notepad via Windows 11) left-clicking and then pasting or using the "Ctrl + V" key-combination after left-clicking. This is rather cumbersome and tedious in the sense of efficacy with respect to the users resource of time (Avoiding philosophical subtleties with respect to the complexity classes and dichotomy in the sense of Kolmogorov & also the power tower or sequential binary operation, to provide an errant opinion in the mode of assumption). Perhaps accessing this page from within the Mathematica console or the Wolfram|One program such as what was mentioned earlier with external APIs in some way navigating a directory safely and in a maximally efficient manner with respect to augmentations such as keypress (Again, lacking the lingo or appropriate terminology with respect to this programming language or the sequence of expressions in the sense of algorithm).

One question is about accessing this course via BigMarker through a secure channel or learning about ports and other nuanced aspects in the sense of being conservative with data and being able to readily upload information through the WDF or contribute in real-time to the Wolfram Function Repository without removing attention from the lecture such as with alternate channels of data in the sense of active execution for formatting and the like.

Hopefully, I can provide less loquacious or verbose inquiries, yet the actual procedure of the computer program (I assume algorithm is correct in the sense of encapsulating it lexically or calling it by one word, as I imagine that I may do if there is audio programming and "calling" the server/use of the kernel involves univocal definitions and sensitivity in terms of attenuation) is what I'm curious about and I've deviated from solving the puzzles or producing a solution set, to my own detriment.

Given that so much time is spent with low-level calculations in Mathematica such as the above, a curiosity is with more compact functions such as with the pure function or defining the concept parametrically (Only having a vague notion of that at present, in the sense of a region that may be implicit, explicit, or parametric, according with abstract or universal geometry [rudimentary mathematical constructions and later demonstrations is a crucial project at present with developing thinking methodically or discerning the methodology, which I think that the Documentation Center does a good job of as it is worded, that I've noticed up until now, although one curiosity, further still, tangentially, is about the LinkSnooper and modifying the formatting with notebooks for stylesheets and whatnot. CloudDeploy[] was noted at 0713 0750 0203 and APIFunction[] on 0863 that date, although I currently lack knowledge of their utility with respect to rough outline of the aforementioned project, failin to attend/being truent for the course on May 16, 2023 on Programming Fundamentals. There is a wonder from 11439 about notation and Boolean Algebra with nn*ee, b^^nnnn, &&, ||, !, and <> about Element[] for the atomic expressions and use of Distributed[], UndirectedEdge[], DirectedEdge[] for Galois Theory (Which was, FiniteField[] and the related group theoretic constructs, experimental as of version 13.3.0) with there a curiosity about Polynomial Algebra and the S- and K-combinatorics with the Chief Executive Officer expounding work from December 1920 (Find WikipediaData@"Combinatory Logic" for a more concise elaboration) and use of bit-vectors in the aforementioned sense of a directory and efficacious navigation through WolframScript via the command line without a graphical interface.

738904.0199

POSTED BY: Updating Name

Hey Arben,

I've not used associations before and I am trying to imagine what is their superpower. I'm not really a software writer, I just play one when I'm trying to study or solve engineering problems. One idea that jumps off the page is using a list of variable value assignments that can be summoned in a program without actually defining the variables outside of the context of the operation using them. That's kind of neat. I'd have to try that out and see if it makes a program easier to write or understand. But I suspect that's just the tip of the iceberg (as is everything else in the Wolfram language).

What do you say is the superpower in using associations?

POSTED BY: Carl Hahn

Hey Carl! Indeed, you can think of associations as constituting some "local map" between keys and values (and indeed, that might be part of the reason why they can be "indexed" with single brackets like functions can, though this is just a wild guess).

The primary value of associations in my view comes from the the organizational structure. For example, suppose I have a nested association:

scores=<|"Bob"-><|"Geography"->100,"Mathematics"->80,"Chemistry"->90|>,
"Alice"-><|"Geography"->90,"Mathematics"->100,"Chemistry"->85|>,
"Carol"-><|"Geography"->Missing["Absent"],"Mathematics"->100,"Chemistry"->95|>|>,

Now, without needing to remember which student or score is where, I can do something like:

scores["Bob"]

or:

scores["Bob","Chemistry"]

or:

scores[[All,"Mathematics"]]

This can be very useful for obvious reasons, especially as datasets get larger and more complicated.

We'll discuss this more later in the course, but you may recall that pure functions work with numbered, unnamed slots: #1+#2&[a,b] returns: a+b But this isn't the only way they can work—they can work with slots named by keys as well: #cat+#dog&[<|"cat"->1,"dog"->2|>] will return 3. This lets you write code that not only extracts the appropriate parts very easy, but performs various queries in a super readable and simple way. For example, the following code will select any students who have a chemistry score of 90 or higher: Select[scores,#Chemistry≥90&] Finally, from a technical perspective, it is faster to add or remove key-value pairs to an association than to add or remove elements with a list. This doesn't really crop up at the example scales we're talking about here, but can have an effect at larger sizes.

POSTED BY: Arben Kalziqi
Posted 11 months ago

Hi arben, Yes this is the tip of iceberg. I suggest you hash table as something to know. O(1) as a search algorithm. Hash tree and topology is a rich context to know

POSTED BY: Mauro Bertani

Hi everybody—I've added today's "filled-out" notebook to the materials folder. See you tomorrow!

POSTED BY: Arben Kalziqi

The link in the email I received doesn't work. How do I navigate to the materials folder? I looked around and didn't see an obvious link.

POSTED BY: David Snyder

Hi David, Apologies for the bad link in today's recording email. The one below should take you right to the folder: https://wolfr.am/WSG48_Programming_Profeciency

POSTED BY: Cassidy Hinkle
Posted 11 months ago

Arben, how do I post without replying to someone else's post (like I am doing now)?

I do have a question about Riffle[ ].

Riffle[{1,2,3}, {x, y}]

outputs {1,x,2,y,3}, understandably. But

Riffle[{x,y}, {1,2,3}]

outputs {x,1,y}. Why? I would have expected to get {x, 1, y, 2}.

Thanks.

POSTED BY: Wissam Barakat

If you want to create a new question on Wolfram Community, please use the button "Start a new Discussion" on the right side of Community's homepage.

enter image description here

Then use the instructions here: http://wolfr.am/READ-1ST

POSTED BY: Ahmed Elbanna

Daily recordings will be shared with the group each day, so you can catch up on any sessions you miss. Looking forward to it!

POSTED BY: Jamie Peterson

I can't make it at this time on MWF, but can on TTh.

POSTED BY: David Snyder

Day 7: Found the notebook to download but not the "data" file. Where can I get this?

POSTED BY: Michael Ulrey

Almost missed this down at the bottom of the thread, but the amoeba folder should now have a data folder there!

POSTED BY: Arben Kalziqi
Reply to this discussion
Community posts can be styled and formatted using the Markdown syntax.
Reply Preview
Attachments
Remove
or Discard

Group Abstract Group Abstract