refactor: Add ElementExpr for _eval expressions
#25199
Conversation
kdn36
requested review from
MarcoGorelli,
alexander-beedie,
c-peters,
orlp,
reswqa and
ritchie46
as code owners
github-actions
bot
added
internal
Codecov Report❌ Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #25199 +/- ##
==========================================
- Coverage 81.81% 81.79% -0.02%
==========================================
Files 1709 1712 +3
Lines 236261 236496 +235
Branches 3005 3007 +2
==========================================
+ Hits 193292 193440 +148
- Misses 42201 42287 +86
- Partials 768 769 +1 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
|
|
||
| out = df.select(pl.col.a.list.eval(expr).over(42)) | ||
| print(out) | ||
| print(df) |
There was a problem hiding this comment.
Remove stray prints
| branch_idx: self.branch_idx, | ||
| flags: self.flags.clone(), | ||
| ext_contexts: self.ext_contexts.clone(), | ||
| element: self.element.clone(), |
There was a problem hiding this comment.
Are we certain this should clone here? Otherwise I think we should be conservative and clear it.
There was a problem hiding this comment.
This would lead to problems, as .split() is also called in BinaryExpr and TernaryExpr and thus would clear the element for different inputs.
There was a problem hiding this comment.
I see. We should add a test case for this then, I imagine it's hit by something like list.eval(pl.when(pl.element()..).then(pl.element()..))
3 participants
This PR introduces
ElementExprand extends the in-memory engine physicalExecutionStatesuch that any input, once evaluated, is retained forpl.element().