This is an automated email from the ASF dual-hosted git repository.
apitrou pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/arrow.git
The following commit(s) were added to refs/heads/main by this push:
new d87bc99de2 GH-45582: [Python] Preserve decimal32/64/256 metadata in
Schema.metadata (#45583)
d87bc99de2 is described below
commit d87bc99de2b7068b42cd57054a4f2d5f21c8d88a
Author: Matthew Roeschke <[email protected]>
AuthorDate: Thu Feb 27 01:53:04 2025 -0800
GH-45582: [Python] Preserve decimal32/64/256 metadata in Schema.metadata
(#45583)
### Rationale for this change
Before, these types would be interpreted as `"object"` type and therefore
the `precision` and `scale` attributes of these types would not be preserved in
the `"metadata"`
### What changes are included in this PR?
Uses `pa.types.is_decimal` is instead of `isinstance`ing just the
`Decimal128Type` to determine a `"decimal"` pandas type
### Are these changes tested?
Yes
### Are there any user-facing changes?
Yes, but should not break compatibility.
* GitHub Issue: #45582
Authored-by: Matthew Roeschke <[email protected]>
Signed-off-by: Antoine Pitrou <[email protected]>
---
python/pyarrow/pandas_compat.py | 2 +-
python/pyarrow/tests/test_pandas.py | 14 ++++++++++++++
2 files changed, 15 insertions(+), 1 deletion(-)
diff --git a/python/pyarrow/pandas_compat.py b/python/pyarrow/pandas_compat.py
index e9655914ad..83f9fd4afa 100644
--- a/python/pyarrow/pandas_compat.py
+++ b/python/pyarrow/pandas_compat.py
@@ -84,7 +84,7 @@ def get_logical_type(arrow_type):
return 'list[{}]'.format(get_logical_type(arrow_type.value_type))
elif isinstance(arrow_type, pa.lib.TimestampType):
return 'datetimetz' if arrow_type.tz is not None else 'datetime'
- elif isinstance(arrow_type, pa.lib.Decimal128Type):
+ elif pa.types.is_decimal(arrow_type):
return 'decimal'
return 'object'
diff --git a/python/pyarrow/tests/test_pandas.py
b/python/pyarrow/tests/test_pandas.py
index 54b89b45f2..7c3ee5ff35 100644
--- a/python/pyarrow/tests/test_pandas.py
+++ b/python/pyarrow/tests/test_pandas.py
@@ -598,6 +598,20 @@ class TestConvertMetadata:
assert data_column['numpy_type'] == 'object'
assert data_column['metadata'] == {'precision': 26, 'scale': 11}
+ @pytest.mark.parametrize('typ', [
+ pa.decimal32,
+ pa.decimal64,
+ pa.decimal128,
+ pa.decimal256,
+ ])
+ def test_decimal_other_bitwidts(self, typ):
+ df = pd.DataFrame({'a': [decimal.Decimal('3.14')]})
+ schema = pa.schema([pa.field('a', type=typ(4, 2))])
+ table = pa.Table.from_pandas(df, schema=schema)
+ col_meta = table.schema.pandas_metadata['columns'][0]
+ assert col_meta['pandas_type'] == 'decimal'
+ assert col_meta['metadata'] == {'precision': 4, 'scale': 2}
+
def test_table_column_subset_metadata(self):
# ARROW-1883
# non-default index
