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.

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

fix: Implement abs for Decimal, error on Date/Time/Datetime #13821

Merged
merged 4 commits into from
Jan 19, 2024
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
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: 25 additions & 10 deletions crates/polars-ops/src/series/ops/abs.rs
Original file line number Diff line number Diff line change
Expand Up @@ -11,19 +11,34 @@ where

/// Convert numerical values to their absolute value.
pub fn abs(s: &Series) -> PolarsResult<Series> {
let physical_s = s.to_physical_repr();
use DataType::*;
let out = match physical_s.dtype() {
let out = match s.dtype() {
#[cfg(feature = "dtype-i8")]
Int8 => abs_numeric(physical_s.i8()?).into_series(),
Int8 => abs_numeric(s.i8().unwrap()).into_series(),
#[cfg(feature = "dtype-i16")]
Int16 => abs_numeric(physical_s.i16()?).into_series(),
Int32 => abs_numeric(physical_s.i32()?).into_series(),
Int64 => abs_numeric(physical_s.i64()?).into_series(),
UInt8 | UInt16 | UInt32 | UInt64 => s.clone(),
Float32 => abs_numeric(physical_s.f32()?).into_series(),
Float64 => abs_numeric(physical_s.f64()?).into_series(),
Int16 => abs_numeric(s.i16().unwrap()).into_series(),
Int32 => abs_numeric(s.i32().unwrap()).into_series(),
Int64 => abs_numeric(s.i64().unwrap()).into_series(),
Float32 => abs_numeric(s.f32().unwrap()).into_series(),
Float64 => abs_numeric(s.f64().unwrap()).into_series(),
#[cfg(feature = "dtype-decimal")]
Decimal(_, _) => {
let ca = s.decimal().unwrap();
let precision = ca.precision();
let scale = ca.scale();

let out = abs_numeric(ca.as_ref());
out.into_decimal_unchecked(precision, scale).into_series()
},
#[cfg(feature = "dtype-duration")]
Duration(_) => {
let physical = s.to_physical_repr();
let ca = physical.i64().unwrap();
let out = abs_numeric(ca).into_series();
out.cast(s.dtype())?
},
dt if dt.is_unsigned_integer() => s.clone(),
dt => polars_bail!(opq = abs, dt),
};
out.cast(s.dtype())
Ok(out)
}
7 changes: 0 additions & 7 deletions py-polars/tests/unit/expr/test_exprs.py
Original file line number Diff line number Diff line change
Expand Up @@ -412,13 +412,6 @@ def test_search_sorted() -> None:
assert a.search_sorted(b, side="right").to_list() == [0, 2, 2, 4, 4]


def test_abs_expr() -> None:
df = pl.DataFrame({"x": [-1, 0, 1]})
out = df.select(abs(pl.col("x")))

assert out["x"].to_list() == [1, 0, 1]


def test_logical_boolean() -> None:
# note, cannot use expressions in logical
# boolean context (eg: and/or/not operators)
Expand Down
6 changes: 0 additions & 6 deletions py-polars/tests/unit/functions/test_functions.py
Original file line number Diff line number Diff line change
@@ -1,6 +1,5 @@
from __future__ import annotations

from datetime import timedelta
from typing import TYPE_CHECKING, Any

import numpy as np
Expand Down Expand Up @@ -390,11 +389,6 @@ def test_fill_null_unknown_output_type() -> None:
}


def test_abs_logical_type() -> None:
s = pl.Series([timedelta(hours=1), timedelta(hours=-1)])
assert s.abs().to_list() == [timedelta(hours=1), timedelta(hours=1)]


def test_approx_n_unique() -> None:
df1 = pl.DataFrame({"a": [None, 1, 2], "b": [None, 2, 1]})

Expand Down
110 changes: 110 additions & 0 deletions py-polars/tests/unit/operations/test_abs.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,110 @@
from __future__ import annotations

from datetime import date, timedelta
from decimal import Decimal as D
from typing import cast

import numpy as np
import pytest

import polars as pl
from polars.testing import assert_frame_equal, assert_series_equal


def test_abs() -> None:
# ints
s = pl.Series([1, -2, 3, -4])
assert_series_equal(s.abs(), pl.Series([1, 2, 3, 4]))
assert_series_equal(cast(pl.Series, np.abs(s)), pl.Series([1, 2, 3, 4]))

# floats
s = pl.Series([1.0, -2.0, 3, -4.0])
assert_series_equal(s.abs(), pl.Series([1.0, 2.0, 3.0, 4.0]))
assert_series_equal(cast(pl.Series, np.abs(s)), pl.Series([1.0, 2.0, 3.0, 4.0]))
assert_series_equal(
pl.select(pl.lit(s).abs()).to_series(), pl.Series([1.0, 2.0, 3.0, 4.0])
)


def test_abs_series_duration() -> None:
s = pl.Series([timedelta(hours=1), timedelta(hours=-1)])
assert s.abs().to_list() == [timedelta(hours=1), timedelta(hours=1)]


def test_abs_expr() -> None:
df = pl.DataFrame({"x": [-1, 0, 1]})
out = df.select(abs(pl.col("x")))

assert out["x"].to_list() == [1, 0, 1]


def test_builtin_abs() -> None:
s = pl.Series("s", [-1, 0, 1, None])
assert abs(s).to_list() == [1, 0, 1, None]


@pytest.mark.parametrize(
"dtype", [pl.Int8, pl.Int16, pl.Int32, pl.Int64, pl.Float32, pl.Float64]
)
def test_abs_builtin(dtype: pl.PolarsDataType) -> None:
lf = pl.LazyFrame({"a": [-1, 0, 1, None]}, schema={"a": dtype})
result = lf.select(abs(pl.col("a")))
expected = pl.LazyFrame({"a": [1, 0, 1, None]}, schema={"a": dtype})
assert_frame_equal(result, expected)


def test_abs_method() -> None:
lf = pl.LazyFrame({"a": [-1, 0, 1, None]})
result_op = lf.select(abs(pl.col("a")))
result_method = lf.select(pl.col("a").abs())
assert_frame_equal(result_op, result_method)


def test_abs_decimal() -> None:
lf = pl.LazyFrame({"a": [D("-1.5"), D("0.0"), D("5.0"), None]})
result = lf.select(pl.col("a").abs())
expected = pl.LazyFrame({"a": [D("1.5"), D("0.0"), D("5.0"), None]})
assert_frame_equal(result, expected)


def test_abs_duration() -> None:
lf = pl.LazyFrame({"a": [timedelta(hours=2), timedelta(days=-2), None]})
result = lf.select(pl.col("a").abs())
expected = pl.LazyFrame({"a": [timedelta(hours=2), timedelta(days=2), None]})
assert_frame_equal(result, expected)


def test_abs_overflow() -> None:
df = pl.DataFrame({"a": [-128]}, schema={"a": pl.Int8})
with pytest.raises(pl.PolarsPanicError, match="attempt to negate with overflow"):
df.select(pl.col("a").abs())


def test_abs_unsigned_int() -> None:
df = pl.DataFrame({"a": [1, 2, 3]}, schema={"a": pl.UInt8})
result = df.select(pl.col("a").abs())
assert_frame_equal(result, df)


def test_abs_non_numeric() -> None:
df = pl.DataFrame({"a": ["p", "q", "r"]})
with pytest.raises(
pl.InvalidOperationError, match="`abs` operation not supported for dtype `str`"
):
df.select(pl.col("a").abs())


def test_abs_date() -> None:
df = pl.DataFrame({"date": [date(1960, 1, 1), date(1970, 1, 1), date(1980, 1, 1)]})

with pytest.raises(
pl.InvalidOperationError, match="`abs` operation not supported for dtype `date`"
):
df.select(pl.col("date").abs())


def test_abs_series_builtin() -> None:
s = pl.Series("a", [-1, 0, 1, None])
result = abs(s)
expected = pl.Series("a", [1, 0, 1, None])
assert_series_equal(result, expected)
20 changes: 0 additions & 20 deletions py-polars/tests/unit/series/test_series.py
Original file line number Diff line number Diff line change
Expand Up @@ -1654,21 +1654,6 @@ def test_temporal_comparison(
)


def test_abs() -> None:
# ints
s = pl.Series([1, -2, 3, -4])
assert_series_equal(s.abs(), pl.Series([1, 2, 3, 4]))
assert_series_equal(cast(pl.Series, np.abs(s)), pl.Series([1, 2, 3, 4]))

# floats
s = pl.Series([1.0, -2.0, 3, -4.0])
assert_series_equal(s.abs(), pl.Series([1.0, 2.0, 3.0, 4.0]))
assert_series_equal(cast(pl.Series, np.abs(s)), pl.Series([1.0, 2.0, 3.0, 4.0]))
assert_series_equal(
pl.select(pl.lit(s).abs()).to_series(), pl.Series([1.0, 2.0, 3.0, 4.0])
)


def test_to_dummies() -> None:
s = pl.Series("a", [1, 2, 3])
result = s.to_dummies()
Expand Down Expand Up @@ -2399,11 +2384,6 @@ def test_repr_html(df: pl.DataFrame) -> None:
assert "<table" in html


def test_builtin_abs() -> None:
s = pl.Series("s", [-1, 0, 1, None])
assert abs(s).to_list() == [1, 0, 1, None]


@pytest.mark.parametrize(
("value", "time_unit", "exp", "exp_type"),
[
Expand Down