Correct format for genes/features.tsv read_10x_mtx

My features.tsv is only one column so a line like this:
genes = pd.read_csv(path / f’{prefix}genes.tsv’, header=None, sep=’\t’)
if var_names == ‘gene_symbols’:
var_names = genes[1].values

in scanpy/readwrite.py at master · theislab/scanpy · GitHub
will fail as it doesn’t have two columns so there is no genes[1] only genes[0]
The column is mix of symbol and ID:
ENSMMUG00000000001__CCNF
I don’t know if that’s direct output of a particular platform or the person that uploaded it on GEO formatted it this way, but I would love like to know what is the solution other than manually breaking the columns into two by the separator ‘__’

I resolved this issue by reformatting the features.tsv, I still would like to know more about the standard inputs expected by scanpy; at least for the dataset I have from GEO I also had to transpose the count matrix as well.