fix: Correct wildcard expansion for functions

[siddharth-vi]

Fixes #19007
Fixes #18968

Remove INPUT_WILDCARD_EXPANSION for the following functions -

• list.contains
• list.count_matches
• pl.duration

INPUT_WILDCARD_EXPANSION is meant for functions which can have variadic arguments. The above functions have fixed number of arguments, hence this flag should not be set.

Behaviour -
Before

import polars as pl

df = pl.DataFrame({"a": [1], "b": [2]})
df.select(pl.duration(hours=pl.all()).name.keep() )

shape: (1, 1)
┌──────────────┐
│ a            │
│ ---          │
│ duration[μs] │
╞══════════════╡
│ 1h 2m        │
└──────────────┘


df = pl.DataFrame({"a": [[1, 2]], "b": [[3, 4]]})
df.select(pl.all().list.contains(3))

#InvalidOperationError: `is_in` operation not supported for dtype `list[i64]`


df = pl.DataFrame({"a": [[1, 2]], "b": [[3, 4]]})
print(df.select(pl.all().list.count_matches(3)))

#SchemaError: could not evaluate comparison between series 'a' of dtype: i64 and series '' of dtype: list[i64]

After

df = pl.DataFrame({"a": [1], "b": [2]})
print(df.select(pl.duration(hours=pl.all()).name.keep() ))

shape: (1, 2)
┌──────────────┬──────────────┐
│ a            ┆ b            │
│ ---          ┆ ---          │
│ duration[μs] ┆ duration[μs] │
╞══════════════╪══════════════╡
│ 1h           ┆ 2h           │
└──────────────┴──────────────┘

df = pl.DataFrame({"a": [[1, 2]], "b": [[3, 4]]})
df.select(pl.all().list.contains(3))

shape: (1, 2)
┌───────┬──────┐
│ a     ┆ b    │
│ ---   ┆ ---  │
│ bool  ┆ bool │
╞═══════╪══════╡
│ false ┆ true │
└───────┴──────┘

df = pl.DataFrame({"a": [[1, 2]], "b": [[3, 4]]})
print(df.select(pl.all().list.count_matches(3)))

shape: (1, 2)
┌─────┬─────┐
│ a   ┆ b   │
│ --- ┆ --- │
│ u32 ┆ u32 │
╞═════╪═════╡
│ 0   ┆ 1   │
└─────┴─────┘

More discussion about this can be found in the issue #19007

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 79.94%. Comparing base (f01fb7b) to head (fbc82a2).
Report is 11 commits behind head on main.

Additional details and impacted files

@@             Coverage Diff             @@
##             main   #19449       +/-   ##
===========================================
+ Coverage   60.05%   79.94%   +19.89%     
===========================================
  Files        1533     1534        +1     
  Lines      210940   211049      +109     
  Branches     2442     2444        +2     
===========================================
+ Hits       126680   168732    +42052     
+ Misses      83706    41762    -41944     
- Partials      554      555        +1     

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

[ritchie46]

Thanks, can you add some tests on the python side?

[siddharth-vi]

@ritchie46 Please check now.

[ritchie46]

Thanks!

[cmdlineluser]

Thank you @siddharth-vi

I'm not sure if it's something you want to look into, but I think there were also some other cases where this seems to happen.

• .list.set_* list.set_intersection operates on the wrong columns when multiple columns are selected with pl.col #18795
• .list.concat

The corresponding .arr.* methods also seemed to demonstrate the same behaviour e.g. .arr.count_matches() etc.

[siddharth-vi]

Will check them out @cmdlineluser