Message Boards Message Boards

GROUPS:

Potential bug with TimeSeriesWindow and InfiniteFuture?

Posted 5 months ago
837 Views
|
2 Replies
|
0 Total Likes
|

I've posted on SE, but it seems not have enjoyed any interest, while it may prove to be an important issue for users of TimeSeries functionality

I use Wolfram Mathematica 12.1 and want to subset a TimeSeries, but noticed that TimeSeriesWindow returns last observation even for range that does not intersect with the original TimeSeries.

Here is an example:

TimeSeriesWindow[
   TimeSeries[Transpose[{#, Range@(Length@#)}]],
   {{2022, 1, 1}, InfiniteFuture}] &@(DateRange[{2000, 1}, {2002, 
    1}, {1, "Month"}])

So intuitively this should return nothing as tmin is {2022, 1, 1} for TimeSeriesWindow, but the TimeSeries ends earlier at {2002, 1}.

But it does and returns TimeSeries[{{{2002,1},12.1}}].

While if I use something like {2100,12,1} instead of InfiniteFuture M will correctly return an error stating that there are no values.

So is it just me? Can you reproduce it? What am I doing wrong?

2 Replies
Posted 5 months ago

It is really amazing how little time series issues are relevant to our community :/

Posted 5 months ago

It does appear to be a bug. You should report the issue to Wolfram Support.

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