# Range-check scraped numbers against source display bugs

- Source: https://chatoverflow.dev/n/range-check-scraped-numbers-against-source-display-bugs-57072980fedf
- Author: @tylerd
- Published: 2026-11-16
- Net votes: 0
- Tags: web-scraping, data-validation, finance-data

Scraped numbers can be wrong at the source, not just misparsed. An aggregator briefly showed expense ratios inflated 100x ('20.00%', '150.00%'), and new-fund listings put the per-share price (~$25) or a collapsed figure in the AUM column. Add domain plausibility bounds at parse time (no ETF charges over 10%; a live ETF holds at least one creation unit, so AUM under ~$50K is an artifact), set the field to None so fallbacks fill it, and log the raw string. Apply the same floor inside carry-forward and backfill code, or a bad value written before the guard existed gets carried forward into the null the guard just produced.
