Append a column to a nested dataset

Questions
1

I have the following nested dataset:

6227Untitled.png
6227Untitled.png242×260 2.17 KB

ds = Dataset[
  <|"x" ->
    {<|"col1" -> 4.2, "col2" -> 2.5|>, <|"col1" -> 1.6, 
      "col2" -> 7.4|>, <|"col1" -> 7.1, "col2" -> 3.6|>}
   ,
   "y" ->
    {<|"col1" -> 9.1, "col2" -> 2.8|>, <|"col1" -> 2.7, 
      "col2" -> 5.4|>, <|"col1" -> 5.3, "col2" -> 0.6|>}
   |>
  ]

How can I append a column col3 to the x sub-dataset, so that col3=f[col1], as below:

Untitled2.png
Untitled2.png349×333 10.1 KB

Doing

ds = ds["x", All, <|#, "new" -> f[#col1]|> &]

adds the column but removes information about the “y” key.

2

Here’s one approach that uses a purpose-built function to decide how to add the column.

AddColumn["x", as_] := "x" -> Map[Append[#, "new" -> f[#col1]] &, as];
AddColumn[key_, as_] := key -> as;

ds[KeyValueMap[AddColumn]/*Apply[Association]]

Or, to re-assign ds

ds = ds[KeyValueMap[AddColumn]/*Apply[Association]]
3

You were quite close!, Try:

ds[All, All, <|#, "col3" -> f[#col1]|> &]
4

Thanks, but this appends a column col3 to both key "x" and key "y":

1573Untitled.png
1573Untitled.png428×288 4.81 KB

I want to apply f[#col1] on the values of key "x" and keep my self the option of applying g[#col1] on the values of key "y".

5

This is one way:

ds // MapAt[Append[#, "col3" -> f[#col1]] &, {"x", All}]