[multistage] Fix AVG Return Type#10368
Conversation
eb9699d to
1625c98
Compare
Codecov Report
@@ Coverage Diff @@
## master #10368 +/- ##
============================================
- Coverage 70.32% 68.37% -1.96%
+ Complexity 5940 5937 -3
============================================
Files 2035 2033 -2
Lines 110228 110193 -35
Branches 16748 16751 +3
============================================
- Hits 77520 75343 -2177
- Misses 27275 29490 +2215
+ Partials 5433 5360 -73
Flags with carried forward coverage won't be shown. Click here to find out more.
📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more |
1625c98 to
14fdbe9
Compare
walterddr
left a comment
There was a problem hiding this comment.
can you also add a test in runtime to make sure avg on a big decimal column returns a big decimal value? (a simple 2 big-decimal that taking an avg resulting in a double-overflowed result should be suffice)
@walterddr Oh yeah, I tried adding it earlier. There's a existing issue where none of the aggregation functions work on BIG_DECIMAL columns 😅 . I'll spend some time tomorrow debugging and fixing it. We can incorporate the fix in this PR or merge this PR and create a separate one for BIG_DECIMAL aggregate functions. I'm ok either way. |
ahh. in this case nvm we can fix in follow up PR |
This PR fixess the return type of AVG aggregation function. The discussion/context behind this change is documented in OSS issue #10318
Added planner tests.