Improve Unicode support for MSSQL#6690
Conversation
|
Ready for review @mistercrunch |
Codecov Report
@@ Coverage Diff @@
## master #6690 +/- ##
===========================================
- Coverage 73.42% 56.07% -17.35%
===========================================
Files 75 525 +450
Lines 10084 23229 +13145
Branches 0 2777 +2777
===========================================
+ Hits 7404 13026 +5622
- Misses 2680 9794 +7114
- Partials 0 409 +409
Continue to review full report at Codecov.
|
|
Bad rebase, back to WIP |
|
This should be ok now, unsure why CI failed the cypress tests. @kristw: The |
This reverts commit c44ae61.
|
I just found that it's ignoring unicode string prefix N when using where_in clause in Jinja template Workaround is to manually rewrite it to something like I'm using mssql+pymssql connection |
* Implement unicode where cluases for mssql queries * Add comment about unicode support on sqla 1.3+
* Revert "creating new circular-json safe stringify and replacing one call (apache#6772)" This reverts commit 147bfb6. * Revert "Improve Unicode support for MSSQL (apache#6690)" This reverts commit 313890c. * Revert "Fix uniqueness constraints on tables table (apache#6718)" This reverts commit 04312db.
The upcoming SQLAlchemy 1.3 release introduces N-prefixed Unicode string handling on MSSQL (see sqlalchemy/sqlalchemy@e1b299d ). This PR introduces support for that feature and has been tested on
masterbranch of SQLAlchemy. This is backwards compatible with earlier versions of SQLAlchemy, so doesn't explicitly require 1.3, it just won't handle unicode values correctly on <1.3. Does not impact other db engines, aka minimal risk of regressions. Fixes #6624Before:
After: