ParquetFile.schema.to_arrow_schema() output is truncated with "(... N chars omitted)" for long lines

[SaraPrager]

Describe the bug, including details regarding any error messages, version, and platform.

When running the following code:

import pyarrow.parquet as pq
parquet_file = pq.ParquetFile('<my file with long schema lines>.parquet')
parquet_file.schema.to_arrow_schema()

the output for long schema lines is truncated and replaced with (... N chars omitted).

For example, a schema line like the following will be truncated:

config: list<element: struct<default_value: string, name: string, keywords: list<element: string>, type: s (... 7 chars omitted)

The issue did not occur in pyarrow version 20, but is present in versions 21 and 22.

Component(s)

Parquet

[raulcd]

Thanks for opening the issue, this seems related to this feature:
#46536

When you say the schema is truncated do you mean the string representation on command line or the Schema instance is wrong?

If it's the string representation you should be able to change the element_size_limit on the to_string method.

[SaraPrager]

Thanks for the quick response!
Yes, the issue was only with the string representation from to_string.
Increasing the element_size_limit parameter as you suggested resolved the problem.

[raulcd]

Thanks for opening (and closing) the issue!
Just for the record I've validated the API docs contains the element_size_limit just in case it wasn't documented there (but it is):
https://arrow.apache.org/docs/python/generated/pyarrow.Schema.html#pyarrow.Schema.to_string