*Note*: migrated from original JIRA: https://issues.apache.org/jira/browse/ARROW-11017 Add support for Parquet schema merging so that we can read data sets where some files have additional columns.