GitHub user pitrou edited a comment on the discussion: A new home for pyarrow-stubs?
I wonder if the overloads can be generated instead of written out and maintained manually. But, also, these overloads would only match for data created directly in Python. As soon as you're reading a Parquet or CSV file, for example (or receive data from any other system or library, such as DuckDB, Polars, cuDF or any ADBC-connected source), the schema is inferred from the file itself and the type checker will only infer the base types. So the overloads might not be as useful as you think in real life. GitHub link: https://github.com/apache/arrow/discussions/45919#discussioncomment-14292822 ---- This is an automatically sent email for [email protected]. To unsubscribe, please send an email to: [email protected]
