Replies: 2 comments 5 replies
-
My personal preference is for #2. Its boring but API users don't seem to appreciate surprises. I agree that the number of parameters on the catalog end-points has grown but apart from splitting the timeseries/location catalog into their own end-points there isn't too much we can do about it. Those end-points are very useful, makes sense that people want lots of options and specializations. Recently added Parameter objects should at least make the Controller<->Dao interactions a tiny bit less tedious. I have no idea what other databases support. Is it just Oracle that doesn't do negative lookahead? Does postgres support it? |
Beta Was this translation helpful? Give feedback.
-
Is this just on location-kind? Or would similar filters be needed for other fields? If it was used on IDs I'd prefer having a like argument and a separate not-like argument because you can combine them together. Isn't location kind limited to a fixed set of options? Seems like you could do an exclude by getting the full set of location-kinds and then querying for just the ones you want. |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
Hi all,
This is in relation to REGI-469, a ticket to add support for REGI queries to CDA. The current PR for this work can be found here: #1211
The work for this ticket involves adding support for a location kind exclusion filter. There is already support for location kind
LIKE
regex matching, but the Oracle implementation of regex in the database (POSIX ERE) does not support usingNOT
operators in the passed expression.As a result, there are three possible options for adding this support that have been identified:
!
prepended syntax. This would involve checking regex for a prepended!
and triggering different JOOQ logic to negate the match.^((?!outlet|turbine|project).)*$
against a pattern such as\^\(\(\?\!.*\)\.\)\*\$
to trigger negation via JOOQ.Option 1 would likely be the cleanest solution, but would add custom syntax to the query parameter. It would certainly need to be documented, as this is non-standard behavior for a regex parameter.
Please provide your thoughts and feedback, as a decision on this implementation will be necessary to move forward.
Beta Was this translation helpful? Give feedback.
All reactions