fixtext-data
JSON ↔ CSV Converter and CSV Explorer v1.1 — no more silent data loss
v1.1Three inputs that used to lose cells with no warning now keep everything and say what happened.
- Duplicate headers. A header row of
a,a,a (2)named the second columna (2), the same as the third, and the third overwrote it. Numbering now skips any name that already appears in the header, so those columns becomea,a (3)anda (2). The CSV Explorer shares this naming code and gets the same fix. - Ragged rows. A row with more cells than the header dropped the extras. They are now kept under
Column 3,Column 4and so on, and a warning reports how many rows were wider and where the first one is. - Flatten collisions. With Flatten nested on,
{"a.b":1,"a":{"b":2}}produced a singlea.bcolumn holding2. Both values are kept, the second asa.b (2), with a warning naming the row.
If you exported from one of these inputs before today, it is worth converting it again.