andras/perfect-postcode
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages 2 Wiki Activity Actions

lgtm

Browse source
This commit is contained in:
Andras Schmelczer 2026-07-12 15:03:33 +01:00
parent 982e0cc89c
commit cfaf58dfba
44 changed files with 793 additions and 201 deletions
Download patch file Download diff file Expand all files Collapse all files

2
pipeline/utils/nspl_schema.py
View file

@ -4,7 +4,7 @@ NSPL/NSUL embed the source year in classification-index column names
(e.g. ruc21ind, oac11ind, imd20ind) and ONS bumps those suffixes with each
release. These columns look numeric in early rows but contain string codes
like "UN1" (Unclassified) further down, so they must be forced to String
before scanning — otherwise polars infers Int64 and crashes mid-stream.
before scanning. Otherwise polars infers Int64 and crashes mid-stream.
Hard-coding the year suffix is fragile: polars silently ignores overrides for
columns that don't exist, so a renamed suffix would no-op and reintroduce the