the-other-tim-brown commented on code in PR #17833:
URL: https://github.com/apache/hudi/pull/17833#discussion_r2850272568
##########
hudi-common/src/main/java/org/apache/hudi/common/schema/HoodieSchemaCompatibilityChecker.java:
##########
@@ -382,6 +383,8 @@ private SchemaCompatibilityResult
calculateCompatibility(final HoodieSchema read
return result.mergedWith(typeMismatch(reader, writer, locations));
case RECORD:
return result.mergedWith(typeMismatch(reader, writer, locations));
+ case VARIANT:
+ return result.mergedWith(typeMismatch(reader, writer, locations));
Review Comment:
Let's add some tests for this? For compatibility, will we allow different
shredded values as long as the reader/writer are variants?
##########
hudi-spark-datasource/hudi-spark4-common/src/main/scala/org/apache/spark/sql/adapter/BaseSpark4Adapter.scala:
##########
@@ -196,4 +198,80 @@ abstract class BaseSpark4Adapter extends SparkAdapter with
Logging {
storageConf.getBoolean(SQLConf.CASE_SENSITIVE.key,
sqlConf.caseSensitiveAnalysis),
getRebaseSpec("CORRECTED"))
}
+
+ override def getVariantDataType: Option[DataType] = {
+ Some(VariantType)
+ }
+
+ override def isDataTypeEqualForParquet(requiredType: DataType, fileType:
DataType): Option[Boolean] = {
+ /**
+ * Checks if a StructType is the physical representation of VariantType in
Parquet.
+ * VariantType is stored in Parquet as a struct with two binary fields:
"value" and "metadata".
+ */
+ def isVariantPhysicalSchema(structType: StructType): Boolean = {
+ if (structType.fields.length != 2) {
+ false
+ } else {
+ val fieldMap = structType.fields.map(f => (f.name, f.dataType)).toMap
Review Comment:
Do we want to enforce a strict ordering here?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
