Victor Uriarte created ARROW-2046:
-------------------------------------

             Summary: [Python] Add support for PEP519 - pathlib and similar 
objects
                 Key: ARROW-2046
                 URL: https://issues.apache.org/jira/browse/ARROW-2046
             Project: Apache Arrow
          Issue Type: Improvement
          Components: Python
            Reporter: Victor Uriarte


Currently `pyarrow` doesn't seem to support reading from `pathlib.Path` or 
similar objects. [PEP519|https://www.python.org/dev/peps/pep-0519/] introduced 
`__fspath__` which could be used to transform any `Path` like object to a 
string.

[Pandas|https://github.com/pandas-dev/pandas/blob/a9d8e04ab68f688f899b4164bfa1ac868c9c1c64/pandas/io/common.py#L120-L160]
 has a sample implementation, though I think a simpler implementation of it 
could be used.

 
{code:java}
import pathlib
import pandas as pd
import pyarrow as pa
import pyarrow.parquet as pq
df = pd.DataFrame({
 'Foo': ['A', 'A', 'B', 'B', 'C'],
 'Bar': ['A1', 'A2', 'B2', 'D3', ''],
})
test_dir = pathlib.Path(__file__).parent / 'test'
test_dir.mkdir(parents=True, exist_ok=True)
table = pa.Table.from_pandas(df)
path = test_dir / 'file1.parquet'
# Doesn't work
pq.write_table(table, path)
# Works
pq.write_table(table, str(path))
{code}
 

 

 



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Reply via email to