Kristin Cowalcijk created SEDONA-408:
----------------------------------------
Summary: Set a reasonable default size for RasterUDT
Key: SEDONA-408
URL: https://issues.apache.org/jira/browse/SEDONA-408
Project: Apache Sedona
Issue Type: Improvement
Reporter: Kristin Cowalcijk
Fix For: 1.5.0
The {{defaultSize}} method of {{UserDefinedType}} is used by Spark SQL query
optimizer to decide whether to broadcast the DataFrame or not. For
{{RasterUDT}}, the default value is 100 bytes, which is the default size of
{{BinaryType}}. This is almost always too small for {{RasterUDT}} and will lead
to large raster DataFrames being mistakenly broadcasted. We can override this
method and set a better default size for {{RasterUDT}}. Maybe 512 KB is a
reasonable default value.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
