top | item 34685004

TableChopper: Edit CSVs with Natural Language

8 points| lunaticd | 3 years ago |tablechopper.pages.dev

2 comments

order

yoz|3 years ago

Very nice! I hit some severe bugs but got several queries working that I didn't expect, such as:

   Add a column named "FirstLetter" with the first letter of "name"
Bugs:

1. In a couple of my test CSVs, some columns were completely blank in the view, though the data came through in queries. I suspect it happens with columns that have values too wide for the cell.

2. 'Group by' queries caused a Python exception in the console log: "AttributeError: 'Series' object has no attribute 'columns'"

brianjking|3 years ago

Neat, is this using Pandas + GPT or something? Looks like you're interacting with a DF in some capacity.

Are you able to query and indicate "How many rows have X data" and have it provide me a count? It seems to only filter the data, doesn't give me actual info about what I've asked.