jsfenfen / 990-xml-metadata Goto Github PK

In return.models:
pf_part_vi has two attributes named AppldTESTxAmt that refer to Part VI Line 6c and Part VI Line 11 of the PF respectively.
SkdASpprtdOrgInfrmtn has two attributes named EIN that refer to filer EIN and Part I Line 12g Column (supported org EIN).
I suspect this is actually an issue with the 990 concordance file / variables sheet as the duplicate issue can be found there as well. I wasn't able to get the default models to migrate to MySQL until I changed db_name for one of the two duplicated fields in the variables.csv file that IRSX uses and the corresponding attribute in return.models. The parser runs fine after the fix - I was mostly curious whether you'd encountered this and how you'd addressed it if you had.
(also thanks for this thing it's amazing)

See: Nonprofit-Open-Data-Collective/irs-990-data-issue-tracker#4

Section headings for PF change. @HFAwesomeCharts notes they have a modified version of the underlying irsx metadata files, maybe ask them if this is something that can be shared? IRSx should be updated with these at some point, it's been tough to find time for this work of late... cc / @lecy

Actually apply the changes that @thirawr sent over so that case insensitive dbs like mysql don't break.

Switch to using two columns, something like start_version and end_version where:

• The start_version is the lowest version that the file supports (currently 2013v1 for structured stuff and 2010 for csvs / groups)
• End_version is only populated if a variable has been discontinued. Leaving these out will simplify file diffs (otherwise, every cell must be changed).

If we encounter a variable that begins, and then ends, and then begins again, we'll just assume that it refers to the same thing and not worry about it being gone?