Closed sosna closed 3 months ago
@sosna I made the following changes:
Rest-cheat-sheet:
e.g.
Pagination:
Sort:
~Something that still needs clarification is the following:~
~If the data are sorted by the components in the order stated in the sort clause, then sort=TIME_PERIOD:desc
would do the following: It would first sort in descending order of TIME_PERIOD and then by the other components in a nondeterministic way, e.g.,~
FREQ,DIM1,TIME_PERIOD,OBS_VALUE
A,A,2022,1.0654
A,B,2022,1.0261
A,A,2021,1.0752
A,B,2021,1.0730
~In consequence, the observations will not be grouped by time series. I don't think that this was intended with sort=TIME_PERIOD:desc
, right?~
~This could be solved by saying:~
~The data are sorted first in a nondeterministic way for any component not included in the sort parameter and then in the order of components in the sort parameter. In this way, sort=TIME_PERIOD:asc
would also be the default sorting value if the data has a time dimension.~
~Ok for you?~
@sosna @stratosn @agent96 I have updated the pull request with the changes discussed and agreed this Feb per email, notably adding the feature to use *
in the sort as replacement for all components (incl. TIME_PERIOD). For me, the specs are fully ready. We can have this reviewed and approved by TWG in June meeting next week if ok for you.
Thanks, @dosse. The PR looks good to me!
Fix #50 Fix #151