fix(python): Ensure consistent read_database behaviour with empty ODBC "iter_batches"

[alexander-beedie]

Minor follow-up to #14916.

• Additionally makes the behaviour consistent between ODBC queries and all other connection types when setting iter_batches=True and we have an empty result set (the batch iterator should be empty, vs a single RecordBatch with no rows).
• Adds "empty dataset" test coverage for KùzuDB.

[codecov]

Codecov Report

Attention: Patch coverage is 50.00000% with 3 lines in your changes are missing coverage. Please review.

Project coverage is 81.01%. Comparing base (e1a4179) to head (58ae892).
Report is 8 commits behind head on main.

Files	Patch %	Lines
py-polars/polars/io/database.py	50.00%	3 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #14918      +/-   ##
==========================================
- Coverage   81.01%   81.01%   -0.01%     
==========================================
  Files        1331     1332       +1     
  Lines      172786   172885      +99     
  Branches     2456     2458       +2     
==========================================
+ Hits       139975   140055      +80     
- Misses      32343    32363      +20     
+ Partials      468      467       -1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.