Yang Zhang
Yang Zhang
### Description #### Background Gluten now use Velox codebase's presto function and implement spark specific version if it has same name but different behavior, and Velox already has prefix when...
The suggestions of disable fetching zlib from cpr master branch has a little bit difference compared with cpr 1.10.5 branch, which leads to situations that zlib still be fetched. After...
Spark support `get` function since v3.4.0, the main difference compared with `element_at` is `get` is 0-based index and return null when index out of bounds. Spark ref: https://spark.apache.org/docs/latest/api/sql/index.html#get
## What changes were proposed in this pull request? (Please fill in changes proposed in this fix) (Fixes: \#ISSUE-ID) ## How was this patch tested? (Please explain how this patch...
## What changes were proposed in this pull request? Velox treat missing column as null, will output incorrect results when spark.sql.parquet.mergeSchema enabled. Fixes: #7174 ## How was this patch tested?...
## What changes were proposed in this pull request? Enable `test reading unaligned pages - test all types` for Spark33,34,35. Enable `parquet v2 pages - rle encoding for boolean value...
### Backend VL (Velox) ### Bug description Gluten has upgrade Spark version to v3.5.3 https://github.com/apache/incubator-gluten/pull/7537, but iceberg may has compatibility issue with Spark 3.5.3. https://github.com/apache/incubator-gluten/pull/7537#issuecomment-2428355077 ### Spark version None ###...
### Description ``` java.lang.NoSuchMethodError: org.apache.spark.sql.catalyst.catalog.SessionCatalog.invalidateCachedTable(Lorg/apache/spark/sql/catalyst/QualifiedTableName;)V at org.apache.spark.sql.hudi.command.DropHoodieTableCommand.run(DropHoodieTableCommand.scala:52) at org.apache.spark.sql.hudi.catalog.HoodieCatalog.dropTable(HoodieCatalog.scala:159) at org.apache.spark.sql.execution.datasources.v2.DropTableExec.run(DropTableExec.scala:38) ```