sadpandajoe commented on PR #45435:
URL: https://github.com/apache/superset/pull/45435#issuecomment-6329731707

   Thanks @dhruv-techdev, the SQLite-based test works for me, so you don't need 
to change approach to land this. Here are a few optional options, cheapest 
first:
   
   1. **Pin the expression.** Add a parametrized exact-string assertion on 
`HiveEngineSpec._time_grain_expressions[TimeGrain.WEEK]`, like `test_ibmi.py` 
does. It's a few lines and catches an accidental revert.
   2. **Note the `'u'` assumption.** The `from_unixtime` stand-in returns 
`isoweekday()`, so the test would still pass if some Hive version formatted 
`'u'` differently. A short comment saying this assumes `SimpleDateFormat` 
semantics would make that clear. If you can run the SQL Lab query from the 
description on a live Hive and paste the output and Hive version, that's the 
real evidence.
   3. **Real Hive in CI.** The `test-postgres-hive` job already starts Hive 
2.3.2 from `scripts/databases/hive/docker-compose.yml`, but it only runs tests 
marked `chart_data_flow` or `sql_json_flow`. A small marked test that runs the 
WEEK expression over 14 literal days against that Hive would replace the 
stand-ins with the real functions. It needs a new marker and a workflow edit, 
so only do it if you're interested. A follow-up PR is fine.
   
   One caveat on option 3: the CI Hive is 2.3.2, which uses `SimpleDateFormat`. 
I haven't checked how newer Hive versions handle `'u'`, and if they differ, a 
separate issue would be the place to track it.
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to