pvary commented on a change in pull request #2418:
URL: https://github.com/apache/hive/pull/2418#discussion_r657004864
##########
File path:
iceberg/iceberg-handler/src/main/java/org/apache/iceberg/mr/hive/HiveIcebergMetaHook.java
##########
@@ -262,17 +264,52 @@ public void
preAlterTable(org.apache.hadoop.hive.metastore.api.Table hmsTable, E
if (AlterTableType.ADDCOLS.equals(currentAlterTableOp)) {
Collection<FieldSchema> addedCols =
- HiveSchemaUtil.schemaDifference(hmsTable.getSd().getCols(),
HiveSchemaUtil.convert(icebergTable.schema()));
+ HiveSchemaUtil.newColumns(hmsTable.getSd().getCols(),
HiveSchemaUtil.convert(icebergTable.schema()));
+ if (!addedCols.isEmpty()) {
+ updateSchema = icebergTable.updateSchema();
+ }
+ for (FieldSchema addedCol : addedCols) {
+ updateSchema.addColumn(
+ addedCol.getName(),
+
HiveSchemaUtil.convert(TypeInfoUtils.getTypeInfoFromTypeString(addedCol.getType())),
+ addedCol.getComment()
+ );
+ }
+ } else if (AlterTableType.REPLACE_COLUMNS.equals(currentAlterTableOp)) {
+ List<FieldSchema> icebergSchema =
HiveSchemaUtil.convert(icebergTable.schema());
+ List<FieldSchema> hmsSchema = hmsTable.getSd().getCols();
+ Collection<FieldSchema> droppedCols =
HiveSchemaUtil.newColumns(icebergSchema, hmsSchema);
+ Collection<FieldSchema> addedCols = HiveSchemaUtil.newColumns(hmsSchema,
icebergSchema);
+ Collection<FieldSchema> typeUpdated =
HiveSchemaUtil.updatedTypeColumns(hmsSchema, icebergSchema);
+ Collection<FieldSchema> commentUpdated =
HiveSchemaUtil.updatedCommentColumns(hmsSchema, icebergSchema);
Review comment:
This is wasteful a little bit - I assume we loop through the columns
every time.
Are the methods used elsewhere? Is the code so much cleaner this way that we
are comfortable to pay the price?
--
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.
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
