Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

fix: Allow for date/datetime subclasses (e.g. pd.Timestamp, FreezeGun) in pl.lit #18497

Merged
merged 4 commits into from
Sep 7, 2024
Merged

fix: Allow for date/datetime subclasses (e.g. pd.Timestamp, FreezeGun) in pl.lit #18497

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
633b000
fix: allow for date/datetime subclasses (e.g. pd.Timestamp, FreezeGun…
MarcoGorelli Aug 31, 2024
33c7058
wip
MarcoGorelli Sep 2, 2024
6f76ebd
Merge remote-tracking branch 'upstream/main' into datetime-subclasses
MarcoGorelli Sep 5, 2024
0d1347d
more robust to allow_object
MarcoGorelli Sep 5, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
35 changes: 24 additions & 11 deletions crates/polars-python/src/functions/lazy.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -436,18 +436,31 @@ pub fn lit(value: &Bound<'_, PyAny>, allow_object: bool) -> PyResult<PyExpr> {
) {
let av = py_object_to_any_value(value, true)?;
Ok(Expr::Literal(LiteralValue::try_from(av).unwrap()).into())
} else if allow_object {
let s = Python::with_gil(|py| {
PySeries::new_object(py, "", vec![ObjectValue::from(value.into_py(py))], false).series
});
Ok(dsl::lit(s).into())
} else {
Err(PyTypeError::new_err(format!(
"cannot create expression literal for value of type {}: {}\
\n\nHint: Pass `allow_object=True` to accept any value and create a literal of type Object.",
value.get_type().qualname()?,
value.repr()?
)))
Python::with_gil(|py| {
// One final attempt before erroring. Do we have a date/datetime subclass?
// E.g. pd.Timestamp, or Freezegun.
let datetime_module = PyModule::import_bound(py, "datetime")?;
let datetime_class = datetime_module.getattr("datetime")?;
let date_class = datetime_module.getattr("date")?;
if value.is_instance(&datetime_class)? || value.is_instance(&date_class)? {
let av = py_object_to_any_value(value, true)?;
Ok(Expr::Literal(LiteralValue::try_from(av).unwrap()).into())
Comment on lines +440 to +448
Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pyo3 does have PyDateTime, but it's limited to:

Available on non-Py_LIMITED_API only

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This should come before the allow_object check.

I am thinking aloud here. Why don't we go via AnyValue directly for all branches? This seems like code duplication.

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

yup, thanks for catching that, and sorry for not having spotted it myself

fixed, and I've added a test which would've on my previous commit to make sure this doesn't regress

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why don't we go via AnyValue directly for all branches?

Tried then, but then a float would become

[crates/polars-python/src/functions/lazy.rs:429:9] &newret = Ok(
    PyExpr {
        inner: 2.0,
    },
)

instead of

[crates/polars-python/src/functions/lazy.rs:432:9] &oldret = Ok(
    PyExpr {
        inner: dyn float: 2.0,
    },
)

, and I haven't (yet) dived deeper here

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Ah, yeap.. That's a good reason.

} else if allow_object {
let s = Python::with_gil(|py| {
PySeries::new_object(py, "", vec![ObjectValue::from(value.into_py(py))], false)
.series
});
Ok(dsl::lit(s).into())
} else {
Err(PyTypeError::new_err(format!(
"cannot create expression literal for value of type {}: {}\
\n\nHint: Pass `allow_object=True` to accept any value and create a literal of type Object.",
value.get_type().qualname()?,
value.repr()?
)))
}
})
}
}

Expand Down
14 changes: 14 additions & 0 deletions py-polars/tests/unit/functions/range/test_date_range.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -310,3 +310,17 @@ def test_date_ranges_datetime_input() -> None:
"literal", [[date(2022, 1, 1), date(2022, 1, 2), date(2022, 1, 3)]]
)
assert_series_equal(result, expected)


def test_date_range_with_subclass_18470_18447() -> None:
class MyAmazingDate(date):
pass

class MyAmazingDatetime(datetime):
pass

result = pl.datetime_range(
MyAmazingDate(2020, 1, 1), MyAmazingDatetime(2020, 1, 2), eager=True
)
expected = pl.Series("literal", [datetime(2020, 1, 1), datetime(2020, 1, 2)])
assert_series_equal(result, expected)
26 changes: 25 additions & 1 deletion py-polars/tests/unit/functions/test_lit.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
from __future__ import annotations

import enum
from datetime import datetime, timedelta
from datetime import date, datetime, timedelta
from decimal import Decimal
from typing import TYPE_CHECKING, Any

Expand Down Expand Up @@ -195,3 +195,27 @@ def test_lit_decimal_parametric(s: pl.Series) -> None:

assert df.dtypes[0] == pl.Decimal(None, scale)
assert result == value


def test_lit_datetime_subclass_w_allow_object() -> None:
class MyAmazingDate(date):
pass

class MyAmazingDatetime(datetime):
pass

result = pl.select(
a=pl.lit(MyAmazingDatetime(2020, 1, 1)),
b=pl.lit(MyAmazingDate(2020, 1, 1)),
c=pl.lit(MyAmazingDatetime(2020, 1, 1), allow_object=True),
d=pl.lit(MyAmazingDate(2020, 1, 1), allow_object=True),
)
expected = pl.DataFrame(
{
"a": [datetime(2020, 1, 1)],
"b": [date(2020, 1, 1)],
"c": [datetime(2020, 1, 1)],
"d": [date(2020, 1, 1)],
}
)
assert_frame_equal(result, expected)