Error on shared column names when writing #1335
Conversation
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## main #1335 +/- ##
==========================================
- Coverage 85.65% 83.47% -2.19%
==========================================
Files 34 34
Lines 5460 5465 +5
==========================================
- Hits 4677 4562 -115
- Misses 783 903 +120
Flags with carried forward coverage won't be shown. Click here to find out more.
|
| df.index, index=df.index | ||
| ).equals(df[df.index.name]): | ||
| raise ValueError( | ||
| f"DataFrame.index.name ({df.index.name!r}) is also used by a column " |
There was a problem hiding this comment.
| f"DataFrame.index.name ({df.index.name!r}) is also used by a column " | |
| f"DataFrame.index.name ({repr(df.index.name)}) is also used by a column " |
https://peps.python.org/pep-0498/#s-r-and-a-are-redundant I had to google this tbh.
There was a problem hiding this comment.
I don't think that PEP is saying that we shouldn't use !r it's just saying that there is another way to do it. Plus it's just a point in the discussion.
For this kind of thing, I would say if our formatter doesn't care then it's fine. If we want to enforce style rules, then we'll do it with the formatter.
| @@ -663,10 +663,23 @@ def write_dataframe(f, key, df, _writer, dataset_kwargs=MappingProxyType({})): | |||
| if reserved in df.columns: | |||
| raise ValueError(f"{reserved!r} is a reserved name for dataframe columns.") | |||
| group = f.require_group(key) | |||
| if not df.columns.is_unique: | |||
| duplicates = list(df.columns[df.columns.duplicated()]) | |||
| raise ValueError( | |||
There was a problem hiding this comment.
A nit, but this seems like a "breaking change" since the behavior is changing. I'm not sure how strict things are about that here, though. Should we warn until 0.11.0? Maybe use #1270 to ease things? But I see the point that this is a "bug" in some sense as well.
There was a problem hiding this comment.
I think this PR is throwing an error on malformed input which was allowed through before (leading to more problems). To me, that's much more fixing of a bug than changing behaviour.
Co-authored-by: Isaac Virshup <[email protected]>
AnnDataobject with duplicated.obscolumn names from.h5ad#884