bmschmidt opened a new issue, #14988:
URL: https://github.com/apache/arrow/issues/14988
### Describe the enhancement requested
Thanks to everyone who works on this project.
I want to report what seems to me a less-than-ideal behavior in the python
API. `Schema.get_field_index()` returns -1 to indicate that a given column is
not in the list (C-style), but `Schema.get_field(-1)` returns the last element
in the list (python-style). In concert these can raise opaque bugs that are a
little hard to track down.
Example:
```python
schema = pa.schema([
pa.field('n_legs', pa.int64()),
pa.field('animals', pa.string()),
])
# Returns -1 to indicate that the column doesn't exist
nonexistent_field = schema.get_field_index("n_eyes")
# Happily consumes -1 to return the field for 'animals.'
schema.field(nonexistent_field)
```
I would expect a python library would raise a `ValueError` when you request
the index for a key that doesn't exist. That's my proposal here, I guess. (That
is, schema.get_field_index should behave the same as python's native
`List.index()` method.); I'm also not convinced that negative-style indexing on
schema.field is very desirable, but that's probably it's the sort of feature
that can't be removed. Failing either of these, it might be worth putting a
note in the docs that `Schema.field` and `Schema.get_field_index` don't play as
well together as one might expect.
### Component(s)
Python
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
