From 6903f803041062904a0a3ce37b5f031597cbd0b3 Mon Sep 17 00:00:00 2001
From: GALI PREM SAGAR <sagarprem75@gmail.com>
Date: Tue, 20 Feb 2024 15:50:57 -0600
Subject: [PATCH] Add support for arrow `large_string` in `cudf` (#15093)

This PR adds support for `large_string` type of `arrow` arrays in `cudf`. `cudf` strings column lacks 64 bit offset support and it is WIP: https://github.com/rapidsai/cudf/issues/13733

This workaround is essential because `pandas-2.2+` is now defaulting to `large_string` type for arrow-strings instead of `string` type.: https://github.com/pandas-dev/pandas/pull/56220

This PR fixes all 25 `dask-cudf` failures.

Authors:
  - GALI PREM SAGAR (https://github.com/galipremsagar)

Approvers:
  - Matthew Roeschke (https://github.com/mroeschke)
  - Ashwin Srinath (https://github.com/shwina)

URL: https://github.com/rapidsai/cudf/pull/15093
---
 python/cudf/cudf/core/column/column.py | 7 +++++++
 python/cudf/cudf/tests/test_series.py  | 8 ++++++++
 python/cudf/cudf/utils/dtypes.py       | 2 ++
 3 files changed, 17 insertions(+)

diff --git a/python/cudf/cudf/core/column/column.py b/python/cudf/cudf/core/column/column.py
index f665d83964c..191c55a8a68 100644
--- a/python/cudf/cudf/core/column/column.py
+++ b/python/cudf/cudf/core/column/column.py
@@ -1920,6 +1920,13 @@ def as_column(
         return col
 
     elif isinstance(arbitrary, (pa.Array, pa.ChunkedArray)):
+        if pa.types.is_large_string(arbitrary.type):
+            # Pandas-2.2+: Pandas defaults to `large_string` type
+            # instead of `string` without data-introspection.
+            # Temporary workaround until cudf has native
+            # support for `LARGE_STRING` i.e., 64 bit offsets
+            arbitrary = arbitrary.cast(pa.string())
+
         if pa.types.is_float16(arbitrary.type):
             raise NotImplementedError(
                 "Type casting from `float16` to `float32` is not "
diff --git a/python/cudf/cudf/tests/test_series.py b/python/cudf/cudf/tests/test_series.py
index 252343391be..caf8947e3b0 100644
--- a/python/cudf/cudf/tests/test_series.py
+++ b/python/cudf/cudf/tests/test_series.py
@@ -2700,3 +2700,11 @@ def test_series_dtype_astypes(data):
     result = cudf.Series(data, dtype="float64")
     expected = cudf.Series([1.0, 2.0, 3.0])
     assert_eq(result, expected)
+
+
+def test_series_from_large_string():
+    pa_large_string_array = pa.array(["a", "b", "c"]).cast(pa.large_string())
+    got = cudf.Series(pa_large_string_array)
+    expected = pd.Series(pa_large_string_array)
+
+    assert_eq(expected, got)
diff --git a/python/cudf/cudf/utils/dtypes.py b/python/cudf/cudf/utils/dtypes.py
index 8fa4a230e2c..c8aca94ba19 100644
--- a/python/cudf/cudf/utils/dtypes.py
+++ b/python/cudf/cudf/utils/dtypes.py
@@ -213,6 +213,8 @@ def cudf_dtype_from_pa_type(typ):
         return cudf.core.dtypes.StructDtype.from_arrow(typ)
     elif pa.types.is_decimal(typ):
         return cudf.core.dtypes.Decimal128Dtype.from_arrow(typ)
+    elif pa.types.is_large_string(typ):
+        return cudf.dtype("str")
     else:
         return cudf.api.types.pandas_dtype(typ.to_pandas_dtype())