-
Notifications
You must be signed in to change notification settings - Fork 10
Description
Description
At-large congressional district H5 files are now named with 0-based numbering (e.g., AK-00.h5) instead of the conventional 1-based numbering (AK-01.h5).
Congressional districts are conventionally 1-indexed — North Carolina's 1st district is NC-01, and Alaska's sole at-large district should be AK-01.
Root cause
Commit 5a04c9f ("Fix at-large district geoid mismatch") correctly changed the internal GEOID for at-large districts from XX01 to XX00 to match Census/FIPS conventions and fix data lookup mismatches. However, the display name logic derives the district number directly from the GEOID (cd_int % 100), so at-large districts now produce 00 in filenames.
Affected states
AK, DE, MT, ND, SD, VT, WY, DC — all at-large districts now get XX-00.h5 instead of XX-01.h5.
Proposed fix
Change district_num = cd_int % 100 to district_num = max(cd_int % 100, 1) in the friendly-name construction, at 4 locations:
publish_local_area.py(3 occurrences, lines ~153, ~231, ~330)stacked_dataset_builder.py(1 occurrence, line ~922)
Internal GEOIDs remain unchanged (XX00). Only the display/filename convention is affected.