Find way to filter CSV within given restraints

[liamreeves]

We can expect the following:
a single rectangular region in the CSV, which may be sparse with empty cells
the first column will always be reserved for consumption by the ETL tool
You can expect cordra.object rows at the bottom and multiple varied cordra.* parser.* above the cordra.object rows
You can expect the first 2 rows to be parser.action and parser.group with the following below in any order before the cordra.object rows:
cordra.relation (optional)
cordra.inverseRelation (optional)
parser.ignore (optional)
cordra.schema (required)
cordra.property (required)
For nested array objects, you can expect array elements to be any quantity 1..n (i.e., any number of affiliations and identifiers)
You can expect any number of groups 1..n
You can expect any level of nesting a/b/c = {“a”:{“b”:{“c”:#cellValue#}}}