DEPR: remove Int/Uint/Float64Index from pandas/tests/indexing & pandas/tests/series #50781
Conversation
topper-123
force-pushed
the
remove_int64_uint64_float64_index_indexing_series_tests
branch
from
7e27bf4 to
638931a
Compare
| @@ -4864,11 +4865,11 @@ def _convert_index(name: str, index: Index, encoding: str, errors: str) -> Index | |||
| atom = DataIndexableCol._get_atom(converted) | |||
|
|
|||
| if ( | |||
| isinstance(index, Int64Index) | |||
| (isinstance(index, NumericIndex) and is_integer_dtype(index)) | |||
There was a problem hiding this comment.
do we need to exclude unsigned here? or non-64-bit?
There was a problem hiding this comment.
It actually makes no difference AFAIKS, because below in lines 4911:4914 there is a section for integers and floats, which returns the same IndexCol (because getattr(index, "freq", None) is None and the same for tz).
So I could either delete this line, or the section on line 4911-4914 or keep both, the result is the same. My opinion is that it would be cleanest to delete this line and handle numeric indexes in the section in lines 4911-4914.
|
Thanks @topper-123 |
topper-123
deleted the
remove_int64_uint64_float64_index_indexing_series_tests
branch
|
Great, It's getting close now to removing Int64Index & friends. Only outstanding big issue is arithemetic with indexes with bit-size lower than 64-bit. I hope I can get that done tomorrow or saturday. |
…s/tests/series (pandas-dev#50781) * remove Int/Uint/Float64Index from pandas/tests/indexing & pandas/tests/series * fix failures * fix failures II * fix failures III * make positive integers explicit Co-authored-by: Terji Petersen <[email protected]>
Extraction of part of #50479. Progress towards #42717.