jorisvandenbossche commented on a change in pull request #12311:
URL: https://github.com/apache/arrow/pull/12311#discussion_r796623200
##########
File path: python/pyarrow/table.pxi
##########
@@ -1771,6 +1771,22 @@ cdef class Table(_PandasConvertible):
raise TypeError(type(item))
result = pyarrow_wrap_table(CTable.Make(c_schema, columns))
+
+ # In case of an empty dataframe with RangeIndex -> create an empty
Table with
+ # number of rows equal to Index length
+ if arrays == [] and schema is not None:
+ try:
+ kind = schema.pandas_metadata["index_columns"][0]["kind"]
+ if kind =="range":
+ start = schema.pandas_metadata["index_columns"][0]["start"]
+ stop = schema.pandas_metadata["index_columns"][0]["stop"]
+ step = schema.pandas_metadata["index_columns"][0]["step"]
+ n_rows = (stop - start - 1)//step + 1
+ result = pyarrow_wrap_table(
+ CTable.MakeWithRows(c_schema, columns, n_rows))
+ except IndexError:
+ pass
Review comment:
Or we could maybe even move the actual check to `dataframe_to_arrays` in
pandas_compat.py, since you already have a `pandas_metadata` object there. And
then we can add a third return argument that is always None, except if you have
this case of no columns but a range index (and then can return the range index
info or length)
(it splits the logic a bit into two places, but it also keeps the very
pandas-specific things in pandas_compat.py)
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
