Clojurians Log v2
sql
Thank you for testing that @antonmos
I'll cut 1.1.x later today!
1π1β€οΈ@seancorfield question about :auto-commit false
in https://github.com/seancorfield/next-jdbc/blob/develop/doc/tips-and-tricks.md#streaming-result-sets-1 you say that it is required to stream results sets.
We just tried it without that and it still seems to be working. Are there edge cases that are not apparent?
Streaming results is inherently database-specific @antonmos but also hasn't changed recently.
Mostly I wanted to check that the "final" version of fold-over-plan is still working for you.
Some DBs require auto-commit false, some don't. Some require fetch-size set to a negative value, some to a positive value. Some require specific concurrency and/or cursor settings, some don't. JDBC is a giant mess in terms of implementation, even tho' it's great as an API in theory.
(you should see all of the DB-specific conditionals scattered throughout the next.jdbc test suite!)
we are on postgres and it seems to be working without auto-commit
will let you know if it keeps working or blows up later π Β―\(γ)/Β―
Interesting. Are you certain that you are indeed seeing streaming behavior?
Also, which version of PG are you using? Perhaps they've improved how PG does streaming (or how the JDBC driver does streaming) in more recent versions that than recommendation was required for? As I've said a few times, I don't use PG (except in a very limited way for testing c.j.j and next.jdbc) so I have to rely on what PG users tell me has been needed for them to achieve the behavior they want...
seancorfield/next.jdbc {:mvn/version "1.1.547"} -- the first 1.1 release! Supports multiple result sets, (parallel) foldable plan results, nested transaction control, find-by-keys enhancements (column selection, basic pagination), etc https://github.com/seancorfield/next-jdbc/releases/tag/v1.1.547 (not news to anyone here -- just making it official).