returns a dataset where the results are grouped by both the operation and the _time column. So like this:
operation _time avg(balance)
a 2021-10-31 17:34:36.0 123
b 2021-11-08 21:38:03.0 -123
The avg(balance) is calculated for each _time value and for each operation value. It doesn't make a lot of sense here because the times are so unique that there is probably not multiple operations in the same timestamp.
Expected behavior
The second argument to the by instruction is actually a t_column_Parameter that behaves differently from the first argument which is a t_row_Parameter. It should pivot the dataset so that each unique value in that column are now the new columns in the dataset, like this:
_time a b
2021-10-31 17:34:36.0 123
2021-11-08 21:38:03.0 -123
Describe the bug
The query:
returns a dataset where the results are grouped by both the operation and the _time column. So like this:
The avg(balance) is calculated for each _time value and for each operation value. It doesn't make a lot of sense here because the times are so unique that there is probably not multiple operations in the same timestamp.
Expected behavior
The second argument to the by instruction is actually a
t_column_Parameterthat behaves differently from the first argument which is at_row_Parameter. It should pivot the dataset so that each unique value in that column are now the new columns in the dataset, like this:How to reproduce
Run the first query in the issue.
Screenshots
Software version
pth_03 version: 5.11.0 pth_06 version: 2.3.0-30-g2cbf0768 pth_10 version: 5.3.0-13-g4a1ff44f
Desktop (please complete the following information if relevant):
Additional context
The number of columns created for the unique values of
t_column_Parametershould be limited to 10.