I am trying to read cellranger count matrix using scanpy. The folder 100PDXT contains matrix.mtx.gz, barcodes.tsv.gz, and features.tsv.gz

100PDXT=sc.read_10x_mtx("~/100PDXT/")

when I call 100PDXT.obs, I see cell names, but the shape is 3153 rows × 0 columns.

AAACCCAAGACCATAA-1
AAACCCACAAGACAAT-1
AAACCCACAGGTCAGA-1
AAACGAAAGCCAGTAG-1
AAACGAAAGTCACAGG-1
...
TTTGGTTGTCGTCAGC-1
TTTGGTTTCTCGGTAA-1
TTTGTTGCACAGTCAT-1
TTTGTTGTCGAAATCC-1

3153 rows × 0 columns

There should be 1 column. From what I have seen in tutorials, the anndata.obs should look like this:

0   AAACCCAAGACCATAA-1
1   AAACCCAAGACCATAA-1
2   AAACCCAAGACCATAC-1
3   AAACCCAAGACGATAT-1
4   AAACCCAAGACCATAT-1
... ...
3149    TTTGGTTGTCGTCACC-1
3150    TTTGGTTGTCGTCAGG-1
3151    TTTGGTTGTCCTCAGC-1
3152    TTTGGTTGTCGTCAGC-1
3153    TTTGGTTGTCATCAGC-1

3153 rows × 1 columns

I need it to have an actual column to be able to add some metadata to 100PDXT.obs, but as it has 0 columns, I can't do anything with it.

How to read the count matrix properly, so the anndata.obs has an actual column of cell names?