v.21.9Improvement

Added output_format_avro_string_column_pattern setting for Avro string columns

Added output_format_avro_string_column_pattern setting to put specified String columns to Avro as string instead of default bytes. Implements #22414. #26245 (Ilya Golshtein).
Added the output_format_avro_string_column_pattern setting to export specified String columns in Avro format as strings instead of the default bytes representation.

Why it matters

This feature addresses the issue where String columns exported to Avro format are encoded as bytes by default, which can be inconvenient or incompatible for certain consumers expecting Avro strings. It improves usability and interoperability of Avro exports by allowing users to specify which String columns should be exported as Avro strings.

How to use it

To enable this feature, set the output_format_avro_string_column_pattern setting with a pattern matching the String column names that should be exported as Avro strings. For example:

SET output_format_avro_string_column_pattern = 'column_name_pattern';


This will cause matching String columns to be serialized as Avro strings instead of bytes in the Avro output format.