From 0b0a914a868fcf8d3d0ff652c13b5cbe4db9f61a Mon Sep 17 00:00:00 2001
From: Gijs Burghoorn <me@gburghoorn.com>
Date: Sun, 3 Nov 2024 18:26:16 +0100
Subject: [PATCH] fix: Fix mask and validity confusion in Parquet String
 decoding (#19614)

---
 .../src/arrow/read/deserialize/binview.rs     |  3 +-
 py-polars/tests/unit/io/test_parquet.py       | 49 ++++++++++++++++++-
 2 files changed, 49 insertions(+), 3 deletions(-)

diff --git a/crates/polars-parquet/src/arrow/read/deserialize/binview.rs b/crates/polars-parquet/src/arrow/read/deserialize/binview.rs
index 62a0e7f2df4e..f9b553bea397 100644
--- a/crates/polars-parquet/src/arrow/read/deserialize/binview.rs
+++ b/crates/polars-parquet/src/arrow/read/deserialize/binview.rs
@@ -276,13 +276,12 @@ fn decode_masked_optional_plain(
             verify_utf8,
         );
     }
-
     if page_validity.unset_bits() == 0 {
         return decode_masked_required_plain(
             num_expected_values,
             values,
             target,
-            page_validity,
+            mask,
             verify_utf8,
         );
     }
diff --git a/py-polars/tests/unit/io/test_parquet.py b/py-polars/tests/unit/io/test_parquet.py
index fe75fccd1ad0..773862591f0e 100644
--- a/py-polars/tests/unit/io/test_parquet.py
+++ b/py-polars/tests/unit/io/test_parquet.py
@@ -20,6 +20,7 @@
 from polars.exceptions import ComputeError
 from polars.testing import assert_frame_equal, assert_series_equal
 from polars.testing.parametric import column, dataframes
+from polars.testing.parametric.strategies.core import series
 
 if TYPE_CHECKING:
     from pathlib import Path
@@ -1560,6 +1561,53 @@ def test_predicate_filtering(
     assert_frame_equal(result, df.filter(expr))
 
 
+@pytest.mark.parametrize(
+    "use_dictionary",
+    [False, True],
+)
+@pytest.mark.parametrize(
+    "data_page_size",
+    [1, None],
+)
+@given(
+    s=series(
+        min_size=1,
+        max_size=10,
+        excluded_dtypes=[
+            pl.Decimal,
+            pl.Categorical,
+            pl.Enum,
+            pl.Struct,  # See #19612.
+        ],
+    ),
+    offset=st.integers(0, 10),
+    length=st.integers(0, 10),
+)
+def test_pyarrow_slice_roundtrip(
+    s: pl.Series,
+    use_dictionary: bool,
+    data_page_size: int | None,
+    offset: int,
+    length: int,
+) -> None:
+    offset %= len(s) + 1
+    length %= len(s) - offset + 1
+
+    f = io.BytesIO()
+    df = s.to_frame()
+    pq.write_table(
+        df.to_arrow(),
+        f,
+        compression="NONE",
+        use_dictionary=use_dictionary,
+        data_page_size=data_page_size,
+    )
+
+    f.seek(0)
+    scanned = pl.scan_parquet(f).slice(offset, length).collect()
+    assert_frame_equal(scanned, df.slice(offset, length))
+
+
 @given(
     df=dataframes(
         min_size=1,
@@ -1579,7 +1627,6 @@ def test_slice_roundtrip(df: pl.DataFrame, offset: int, length: int) -> None:
     df.write_parquet(f)
 
     f.seek(0)
-    print((offset, length))
     scanned = pl.scan_parquet(f).slice(offset, length).collect()
     assert_frame_equal(scanned, df.slice(offset, length))