-
Notifications
You must be signed in to change notification settings - Fork 1.5k
Closed
Description
The parquet-avro object model currently only uses String for binary/utf8 data. It should use Utf8, which doesn't copy data and require 2x the space whenever Avro would. That is by default in generic, unless avro.java.string is set to "string". Avro reflect always uses String, and both reflect and specific will respect the java-class property for stringable types.
Reporter: Ryan Blue / @rdblue
Assignee: Ryan Blue / @rdblue
Related issues:
- Release Parquet 1.8.0 (blocks)
- Add avro-reflect support (depends upon)
PRs and other links:
Note: This issue was originally created as PARQUET-286. Please see the migration documentation for further details.