Add compatibility with time_agg with only one operand

[javihern98]

Overview

Currently we support the time aggregation when we pass the dataset. Seems there is another option where we do not pass any operand to the time_agg (currently not supported):

DS_r := sum ( DS_1 group all time_agg ( "A" ) )

Error shown when generating the AST:

vtlengine.Exceptions.SemanticError: ('Optional or empty expression node is not allowed in time_agg.', '1-4-2-2')

Tasks to perform

• Review the manual and discuss the potential behaviour of this function
• Add a ticket to the VTL community if necessary
• Implement the expected behaviour and remove this Semantic Error

[javihern98]

@antonio-olleros Please review this (not urgent)

[antonio-olleros]

Hi Javier, how do we support it? Do we need to add the time component code? If that's the case, then I would say we need to support also this case, with a check: The dataset on which we are aggregating needs to have one, and only one, time dimension

[javihern98]

Currently we add the time component code in the expression. I agree on the check.

[mla2001]

It is not necessary to indicate the time id, there is a method that detects it from a dataset. Check the following method:
Time.py: 60
_get_time_id(cls, operand: Dataset)