promptdojo_

Mission: API rows to a model-ready DataFrame — step 5 of 7

This cleaner swallows bad records with a bare pass — records vanish with no count (the ch.03 sin, pipeline edition). Fix it to COUNT the drops so the report prints kept: 2 dropped: 1.

The break is on line 8 — but read the whole snippet first.

full-screen editor opens — close anytime to keep reading.