Keyword-only key argument for pd.DataFrame.to_hdf() - #3185
Conversation
Sorry, something went wrong.
This keyword-only usage is required for pandas 2.2 and higher.
There was a problem hiding this comment.
Pull request overview
This PR updates pd.DataFrame.to_hdf() calls across multiple files to make the key parameter keyword-only, ensuring compatibility with Pandas 3.0 (released January 2026). This change addresses deprecation warnings in Pandas 2.2+ and errors in Pandas 3.0 where positional usage of the key argument is no longer supported.
Changes:
- Updated all remaining instances of df.to_hdf(path, "key_name", ...) to df.to_hdf(path, key="key_name", ...)
- Applied changes consistently across test scripts, examples, and core library modules
- Ensured compatibility with Pandas 3.0 API requirements
Reviewed changes
Copilot reviewed 12 out of 12 changed files in this pull request and generated no comments.
Show a summary per file| File | Description |
|---|---|
| testscript_cli.py | Updated to_hdf() call to use keyword-only key parameter |
| examples/testscript_transreid.py | Updated two to_hdf() calls to use keyword-only key parameter |
| examples/testscript_pretrained_models.py | Updated two to_hdf() calls to use keyword-only key parameter |
| examples/testscript_multianimal.py | Updated two to_hdf() calls to use keyword-only key parameter |
| examples/testscript_mobilenets.py | Updated two to_hdf() calls to use keyword-only key parameter |
| examples/testscript_deterministicwithResNet152.py | Updated to_hdf() call to use keyword-only key parameter |
| deeplabcut/refine_training_dataset/tracklets.py | Updated to_hdf() call in save method to use keyword-only key parameter |
| deeplabcut/refine_training_dataset/stitch.py | Updated to_hdf() call to use keyword-only key parameter for "tracks" |
| deeplabcut/refine_training_dataset/outlier_frames.py | Updated to_hdf() call to use keyword-only key parameter |
| deeplabcut/post_processing/analyze_skeleton.py | Updated to_hdf() call to use keyword-only key parameter |
| deeplabcut/pose_estimation_tensorflow/core/evaluate.py | Updated to_hdf() call to use keyword-only key parameter |
| deeplabcut/modelzoo/generalized_data_converter/utils.py | Updated to_hdf() call to use keyword-only key parameter |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Sorry, something went wrong.
|
Minimal changes. Should help to prevent issues when users have a wrong pandas installation (also prevented by our restricted requirements, but good to address anyway). |
Sorry, something went wrong.
C-Achard
left a comment
There was a problem hiding this comment.
Should be fine, very low risk
Sorry, something went wrong.
Motivation:
Pandas officially released version 3.0 (January 2026), but our code is not compatible with this version.
Changes:
This PR replaces df.to_hdf(dataname, "df_with_missing", ...) with df.to_hdf(dataname, key="df_with_missing", ...) (and similar) where that was not yet implemented.