ramitg254 commented on code in PR #6565:
URL: https://github.com/apache/hive/pull/6565#discussion_r3503931068
##########
iceberg/iceberg-handler/src/main/java/org/apache/iceberg/mr/hive/HiveIcebergMetaHook.java:
##########
@@ -744,12 +744,14 @@ private void
handleAddColumns(org.apache.hadoop.hive.metastore.api.Table hmsTabl
boolean isORc = isOrcFileFormat(hmsTable);
for (FieldSchema addedCol : addedCols) {
String defaultValue = defaultValues.get(addedCol.getName());
- Type type =
HiveSchemaUtil.convert(TypeInfoUtils.getTypeInfoFromTypeString(addedCol.getType()),
defaultValue);
- Literal<Object> defaultVal = Optional.ofNullable(defaultValue).filter(v
-> !type.isStructType())
- .map(v -> Expressions.lit(HiveSchemaUtil.getDefaultValue(v,
type))).orElse(null);
-
+ Type baseType =
HiveSchemaUtil.convert(TypeInfoUtils.getTypeInfoFromTypeString(addedCol.getType()),
defaultValue);
+ final Type resolvedType = (!isORc && baseType.isStructType()) ?
+ HiveSchemaUtil.applyInitialDefaultsToStruct(baseType) :
+ baseType;
+ Literal<Object> defaultVal = Optional.ofNullable(defaultValue).filter(v
-> !resolvedType.isStructType())
+ .map(v -> Expressions.lit(HiveSchemaUtil.getDefaultValue(v,
resolvedType))).orElse(null);
// ORC doesn't have support for initialDefault from iceberg layer, we
only need to set default for writeDefault.
- updateSchema.addColumn(addedCol.getName(), type, addedCol.getComment(),
isORc ? null : defaultVal);
+ updateSchema.addColumn(addedCol.getName(), resolvedType,
addedCol.getComment(), isORc ? null : defaultVal);
Review Comment:
yes that was the earlier approach I was using but `isOrc` argument needs to
be passed there in convert to determine initial default needs to be set or not
for inner struct fields which doesn't looks good as it will introduce concern
"which metadata should be attached while building column type" which is not
objective of convert , so building a copy of type already built with the
already present metadata and initial defaults of underlying fields if col is of
struct type and isOrc is false
--
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]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
