[
https://issues.apache.org/jira/browse/GOBBLIN-1802?focusedWorklogId=853593&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-853593
]
ASF GitHub Bot logged work on GOBBLIN-1802:
-------------------------------------------
Author: ASF GitHub Bot
Created on: 29/Mar/23 07:33
Start Date: 29/Mar/23 07:33
Worklog Time Spent: 10m
Work Description: phet commented on code in PR #3663:
URL: https://github.com/apache/gobblin/pull/3663#discussion_r1151498974
##########
gobblin-data-management/src/main/java/org/apache/gobblin/data/management/copy/iceberg/IcebergDatasetFinder.java:
##########
@@ -98,20 +107,42 @@ public Iterator<IcebergDataset> getDatasetsIterator()
throws IOException {
return findDatasets().iterator();
}
- protected IcebergDataset createIcebergDataset(String dbName, String tblName,
IcebergCatalog icebergCatalog, Properties properties, FileSystem fs) {
- IcebergTable icebergTable = icebergCatalog.openTable(dbName, tblName);
- return new IcebergDataset(dbName, tblName, icebergTable, properties, fs);
+ /**
+ * Requires both source and destination catalogs to connect to their
respective {@link IcebergTable}
+ * Note: the destination side {@link IcebergTable} should be present before
initiating replication
+ * @return {@link IcebergDataset} with its corresponding source and
destination {@link IcebergTable}
+ */
+ protected IcebergDataset createIcebergDataset(String dbName, String tblName,
IcebergCatalog sourceIcebergCatalog, IcebergCatalog destinationIcebergCatalog,
Properties properties, FileSystem fs) throws IOException {
+ IcebergTable srcIcebergTable = sourceIcebergCatalog.openTable(dbName,
tblName);
+ IcebergTable destIcebergTable =
destinationIcebergCatalog.openTable(dbName, tblName);
+
Preconditions.checkArgument(destinationIcebergCatalog.tableAlreadyExists(TableIdentifier.of(dbName,
tblName)), "Missing Destination Iceberg Table!");
Review Comment:
be sure to name the table in the exception message (to highlight potentially
incorrect config)
##########
gobblin-data-management/src/main/java/org/apache/gobblin/data/management/copy/iceberg/BaseIcebergCatalog.java:
##########
@@ -17,7 +17,9 @@
package org.apache.gobblin.data.management.copy.iceberg;
+import java.io.IOException;
Review Comment:
likely to be a lint error
(FYI in other files too)
##########
gobblin-data-management/src/main/java/org/apache/gobblin/data/management/copy/iceberg/IcebergHiveCatalog.java:
##########
@@ -54,4 +56,9 @@ public String getCatalogUri() {
protected TableOperations createTableOperations(TableIdentifier tableId) {
return hc.newTableOps(tableId);
}
+
+ @Override
+ public boolean tableAlreadyExists(TableIdentifier tableId) {
Review Comment:
I see why it would be difficult to make `tableAlreadyExists()` a method on
`IcebergTable` (as that would require the table to store a pointer back to its
catalog), but why not better encapsulate it, to avoid introducing
`TableIdentifier` into the API? e..g two forms:
* `IcebergCatalog.tableAlreadyExists(String dbName, String tableName)` -
that parallels `openTable`'s signature
* `IcebergCatalog.tableAlreadyExists(IcebergTable table)` - for this, use
your new getter on `IcebergTable.tableId`
Issue Time Tracking
-------------------
Worklog Id: (was: 853593)
Time Spent: 2.5h (was: 2h 20m)
> Register iceberg table metadata update with destination side catalog
> --------------------------------------------------------------------
>
> Key: GOBBLIN-1802
> URL: https://issues.apache.org/jira/browse/GOBBLIN-1802
> Project: Apache Gobblin
> Issue Type: New Feature
> Reporter: Meeth Gala
> Priority: Major
> Time Spent: 2.5h
> Remaining Estimate: 0h
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
