Igor Tavares created ARROW-6805:
-----------------------------------
Summary: Should the new field_ argument on Table.set_column() be
optional? (Unnecessary breaking change)
Key: ARROW-6805
URL: https://issues.apache.org/jira/browse/ARROW-6805
Project: Apache Arrow
Issue Type: Bug
Components: Python
Affects Versions: 0.15.0
Environment: All
Reporter: Igor Tavares
The new *field_* argument (Table.set_column() on version 0.15.0) helps to add
extra information on the column metadata. But it should not be mandatory, in
some cases it is simply redundant like that:
{code:python}
import pandas as pd
import pyarrow as pa
df = pd.DataFrame({"foo": [1, 2, 3]})
tbl = pa.Table.from_pandas(df=df, preserve_index=False)
if pa.__version__ == "0.15.0":
field = pa.field(name="foo", type="double")
tbl = tbl.set_column(0, field, tbl.column("foo").cast("double"))
else:
tbl = tbl.set_column(0, tbl.column("foo").cast("double"))
{code}
I think that this argument should be optional to avoid redundant code and to
keep compatibility with version <0.15.0.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
