- Add support for dask dataframes
- Fix using
S
indf.groupby(...)[col].agg(S. ...)
- Move
C
andI
to stable API: drop warning that they are experimental - Make string representation of
C
andI
even prettier, e.g.,~C['a':'d'] & C.dtype.isin({int, float})
renders exactly the same. - Add helper functions in
paddles
submodule:str_join
,build_filter
,combine
.
- Add reverse binary operators, e.g.
__radd__
.
- Fix using
S
ingroupby().agg()
(GH-21)
- Add
report
function. - Add prettier string representations for
C
andI
expressions.
- Don't warn when creating
C
andI
instances.
- Refactor code: Give DF and S supporting classes better names.
- Allow selecting slices of labels, e.g.,
C["C":"G"]
. - Support
pandas
version 2.
- Add
I
for label-selection for index similar asC
for columns. (Still experimental)
- Fix build of stub package (
pandas_selector
).
- Rename package to
pandas_paddles
. - Drop Python 3.6 from actively supported releases.
- Add Python 3.10 to test matrix.
- Add experimental warning when using
C
.
- Add experimental
C
object to simplify column selection and sorting.
- Add typing hints.
- Raise test-coverage.
- Expand
DF
in wrapped function arguments. - Update Sphinx.
- Fix serialization of
DF
andS
withpickle
.
- Add series accessor (#3)
- Improve doc handling by passing through original pandas documentation for wrapped objects.
- Host documentation on RTD.
- Add project metadata.
- Initial release.
- Add GitHub actions CI pipeline.
- Add test coverage report.
- Inital set of test cases.