0.10.6 - Cannot write files due to implicit conversion to str dtype. #1417
Description
Please make sure these conditions are met
- I have checked that this issue has not already been reported.
- I have confirmed this bug exists on the latest version of anndata.
- (optional) I have confirmed this bug exists on the master branch of anndata.
Report
The new ValueError check introduced in 0.10.6 causes anndata.Anndata to error out on write.
This is because the index is implicitly converted to str by anndata/_core/aligned_df.py:_gen_dataframe_df.
To prevent this, one needs to also do this implicit conversion before performing this check. I have done this in PR #1418 which currently has failing tests and is in draft.
Code:
import anndata
import pandas as pd
df = pd.DataFrame({"column1": [1, 2, 3]})
df.index = df.column1
adata = anndata.AnnData(obs=df)
adata.write_h5ad("cannot_write.h5ad")Traceback:
.../lib/python3.10/site-packages/anndata/_io/specs/methods.py
ValueError: DataFrame.index.name ('column1') is also used by a column whose values are different. This is not supported. Please make sure the values are the same, or use a different name.
Error raised while writing key 'obs' of <class 'h5py._hl.group.Group'> to /Versions
-----
anndata 0.10.6
session_info 1.0.0
-----
astunparse 1.6.3
awkward 2.3.2
awkward_cpp NA
cython_runtime NA
dateutil 2.8.2
dill 0.3.7
exceptiongroup 1.1.2
h5py 3.10.0
mpmath 1.3.0
natsort 8.4.0
numexpr 2.9.0
numpy 1.26.4
packaging 23.1
pandas 2.2.0
pyarrow 15.0.0
pynvml 11.5.0
pytz 2024.1
regex 2.5.140
scipy 1.12.0
setuptools_scm NA
six 1.16.0
sympy 1.12
torch 2.1.0+cu121
torchgen NA
tqdm 4.66.2
typing_extensions NA
zope NA
-----
Python 3.10.4 (main, May 3 2023, 14:08:02) [GCC 11.3.0]
Linux-6.6.10-76060610-generic-x86_64-with-glibc2.35
-----
Session information updated at 2024-03-12 14:21