ENH: Add autofilter parameter to DataFrame.to_excel #62928
Conversation
…ge for xlsxwriter/openpyxl; keep engine_kwargs semantics intact
…ter and a user guide snippet
|
Hi @rhshadrach, I hope you’re doing well. |
There was a problem hiding this comment.
Thanks for the PR! I think this is looking good. Unfortunately it seems that odfpy does not provide a reasonable way to add an autofilter. If that is the case, I think we should raise when autofilter=True is passed and this engine is used. Can you add this and a test.
Can you also add tests for:
- nonzero startrow / startcol
- A DataFrame with MultiIndex columns (also called a hierarchical index) with both
merge_cells=Trueandmerge_cells=False.
| autofilter: bool = False, | ||
| ) -> None: | ||
| """ | ||
| Write object to an Excel sheet. |
There was a problem hiding this comment.
Can you remove the doc decorator on L2149
There was a problem hiding this comment.
Also - it looks like there was a merge issue here and code got duplicated below.
There was a problem hiding this comment.
Can you remove the doc decorator on L2149
@rhshadrach
It’s used for docstring parameter substitution. Removing it would break the docstring formatting used across pandas methods. The autofilter parameter is already documented in the Parameters section (around line 2252):
autofilter : bool, default False
Whether to apply autofilter to the header row.
Is there a specific reason to remove the decorator, or would you prefer a different docstring format? If you saw something else at L2149, can you point me to the exact location?
| # track bounds (1-based for openpyxl) | ||
| if min_r is None or abs_row < min_r: |
There was a problem hiding this comment.
Are we able to determine these by startrow / startcol along with the DataFrame shape?
There was a problem hiding this comment.
Are we able to determine these by
startrow/startcolalong with the DataFrame shape?
@rhshadrach
I looked into calculating the autofilter bounds from startrow/startcol and DataFrame shape, but _write_cells only receives the cells list, startrow, and startcol—not the DataFrame shape, header/index flags, or MultiIndex structure.
To use shape-based calculation, we'd need to:
- Pass additional parameters (DataFrame shape, header/index flags, MultiIndex levels) through the call chain
- Handle edge cases like filtered columns, custom formatting that changes row/col counts, and MultiIndex headers with varying depths
The current approach tracks bounds while iterating through the cells we're already writing:
- Matches what's actually written
- No API changes needed
-Minimal overhead - Handles edge cases (MultiIndex, merged cells, conditional formatting)
If you prefer a shape-based calculation, I can add those parameters to _write_cells, though it will increase complexity. The current tracking approach is straightforward and accurate.
Okay, Im working on them now |
- Remove duplicate to_excel function code in generic.py - Add NotImplementedError for odfpy engine when autofilter=True - Remove broad exception handling from autofilter implementations - Add comprehensive tests for nonzero startrow/startcol - Add tests for MultiIndex columns with merge_cells=True and False - Improve tests to verify each column has autofilter - Remove redundant test_to_excel test - Remove redundant pytest.importorskip from test functions
@rhshadrach
Adds an
autofilterparameter toDataFrame.to_excel()to enable Excel autofilterover the header row and data range when writing files.
Usage:
Supports
xlsxwriterandopenpyxlengines. Whenautofilter=True, applies anautofilter over the written data range (header row through last data row).
Closes #62651