I have varying outcomes with which subset_types I encounter per parse. Although I have not encountered the ones encoded set(['individual', 'disease_state']), I have had to make use of set(['dose', 'agent', 'time', 'gender']). So, in parse_GDS_columns, I modified the code to start out with an empty subset_ids and collected everything on the fly.
This turned out nicely as each subset_type was accounted for in each sample in the end, so in GDS.columns no rows were dropped during GDS.__init__().
I have varying outcomes with which subset_types I encounter per parse. Although I have not encountered the ones encoded
set(['individual', 'disease_state'])
, I have had to make use ofset(['dose', 'agent', 'time', 'gender'])
. So, inparse_GDS_columns
, I modified the code to start out with an emptysubset_ids
and collected everything on the fly.This turned out nicely as each subset_type was accounted for in each sample in the end, so in
GDS.columns
no rows were dropped duringGDS.__init__()
.