Closed philipwfowler closed 1 year ago
This was the result of some weird typing by pandas
. As it requires consistent data typing, for fields such as codon_idx
which weren't always not None, it reverted to using numpy.int64
.
Fixed by converting to pandas
's own nullable integer type Int64
, and filtering before joining to the JSON to convert from pandas
NaN
to pure Python None
values.
Other samples fail with an empty
json
output file. If I try and process the attached vcf I get the following Traceback.site.10.subj.YA00099906.lab.YA00099906.iso.1.v0.12.4.per_sample.vcf.zip minor_alleles.txt