[SEDONA-226] Support reading and writing GeoParquet file metadata

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 1.4.0
• Labels: pull-request-available
• Assignee: Unassigned
• Reporter: Kristin Cowalcijk
• Created: 29/Dec/22 10:55
• Updated: 16/Feb/23 04:05
• Resolved: 04/Jan/23 22:39

Description

Currently, GeoParquetFileFormat in Apache Sedona does not read or write the geo metadata in GeoParquet files. We'll honor geo metadata when loading GeoParquet files and populate geo metadata when writing dataframes to GeoParquet files in order to conform with the standard and improve interoperability with other libraries such as GeoPandas.

We'll drop the fieldGeometry option for loading GeoParquet files since we can infer geometry fields from the geo.columns metadata, there's no need for users to specify the names of geometry fields explicitly.

Issue Links

• is related to: SEDONA-207 Faster serialization/deserialization of geometry objects
• links to: GitHub Pull Request #740
• links to: GitHub Pull Request #771