[
https://issues.apache.org/jira/browse/GOBBLIN-2159?focusedWorklogId=939635&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-939635
]
ASF GitHub Bot logged work on GOBBLIN-2159:
-------------------------------------------
Author: ASF GitHub Bot
Created on: 23/Oct/24 07:48
Start Date: 23/Oct/24 07:48
Worklog Time Spent: 10m
Work Description: phet commented on code in PR #4058:
URL: https://github.com/apache/gobblin/pull/4058#discussion_r1812012723
##########
gobblin-data-management/src/test/java/org/apache/gobblin/data/management/copy/iceberg/IcebergTableTest.java:
##########
@@ -226,6 +226,90 @@ public void testNewTablePropertiesAreRegistered() throws
Exception {
catalog.dropTable(destTableId);
}
+ /** Verify that getPartitionSpecificDataFiles return datafiles belonging to
the partition defined by predicate */
+ @Test
+ public void testGetPartitionSpecificDataFiles() throws IOException {
+ List<String> paths = Arrays.asList(
+ "/path/tableName/data/id=1/file1.orc",
+ "/path/tableName/data/file3.orc",
+ "/path/tableName/data/id=2/file5.orc",
+ "/path/tableName/data/file4.orc",
+ "/path/tableName/data/id=3/file2.orc"
+ );
+ // Using the schema defined in start of this class
+ PartitionData partitionData = new
PartitionData(icebergPartitionSpec.partitionType());
+ partitionData.set(0, "1");
+ Map<String, PartitionData> pathsWithPartitionData = Maps.newHashMap();
+ paths.forEach(path -> pathsWithPartitionData.put(path, partitionData));
+
+ addPartitionDataFiles(table, createDataFiles(pathsWithPartitionData));
+
+ IcebergTable icebergTable = new IcebergTable(tableId,
+ catalog.newTableOps(tableId),
+ catalogUri,
+ catalog.loadTable(tableId));
+ // Using AlwaysTrue & AlwaysFalse Predicate to avoid mocking of predicate
class
+ Predicate<StructLike> alwaysTruePredicate = partition -> true;
+ Predicate<StructLike> alwaysFalsePredicate = partition -> false;
+
Assert.assertEquals(icebergTable.getPartitionSpecificDataFiles(alwaysTruePredicate).size(),
5);
+
Assert.assertEquals(icebergTable.getPartitionSpecificDataFiles(alwaysFalsePredicate).size(),
0);
+ }
+
+ /** Verify that overwritePartition replace data files belonging to given
partition col and value */
+ @Test
+ public void testOverwritePartition() throws IOException {
+ List<String> paths = Arrays.asList(
+ "/path/tableName/data/id=1/file1.orc",
+ "/path/tableName/data/file2.orc"
+ );
+ // Using the schema defined in start of this class
+ PartitionData partitionData = new
PartitionData(icebergPartitionSpec.partitionType());
+ partitionData.set(0, "1");
+ Map<String, PartitionData> pathsWithPartitionData = Maps.newHashMap();
+ paths.forEach(path -> pathsWithPartitionData.put(path, partitionData));
+
+ addPartitionDataFiles(table, createDataFiles(pathsWithPartitionData));
+
+ IcebergTable icebergTable = new IcebergTable(tableId,
+ catalog.newTableOps(tableId),
+ catalogUri,
+ catalog.loadTable(tableId));
+
+ verifyAnyOrder(paths,
icebergTable.getCurrentSnapshotInfo().getAllDataFilePaths(), "data filepaths
should match");
+
+ List<String> paths2 = Arrays.asList(
+ "/path/tableName/data/file3.orc",
+ "/path/tableName/data/id=2/file4.orc"
+ );
+ // Using the schema defined in start of this class
+ PartitionData partitionData2 = new
PartitionData(icebergPartitionSpec.partitionType());
+ partitionData2.set(0, "2");
+ Map<String, PartitionData> paths2WithPartitionData2 = Maps.newHashMap();
+ paths2.forEach(path -> paths2WithPartitionData2.put(path, partitionData2));
+
+ List<DataFile> partition2DataFiles =
createDataFiles(paths2WithPartitionData2);
+ // here, since partition data with value 2 doesn't exist yet,
+ // we expect it to get added to the table, w/o changing or deleting any
other partitions
+ icebergTable.overwritePartition(partition2DataFiles, "id", "2");
+ List<String> expectedPaths2 = new ArrayList<>(paths);
+ expectedPaths2.addAll(paths2);
+ verifyAnyOrder(expectedPaths2,
icebergTable.getCurrentSnapshotInfo().getAllDataFilePaths(), "data filepaths
should match");
+
+ List<String> paths3 = Arrays.asList(
+ "/path/tableName/data/id=2/file5.orc",
+ "/path/tableName/data/file6.orc"
+ );
+ // Reusing same partition data to create data file with different paths
+ Map<String, PartitionData> paths3WithPartitionData = Maps.newHashMap();
+ paths3.forEach(path -> paths3WithPartitionData.put(path, partitionData));
+ List<DataFile> partition1NewDataFiles =
createDataFiles(paths3WithPartitionData);
Review Comment:
NBD, but for a one-liner:
```
List<DataFile> partition1NewDataFiles = createDataFiles(
paths3.stream().collect(Collectors.toMap(x -> x, partition1Data))
);
```
(alternative to `x -> x` is `Function.identity()` - your choice)
Issue Time Tracking
-------------------
Worklog Id: (was: 939635)
Time Spent: 10h 50m (was: 10h 40m)
> Support Partition Based Copy in Iceberg Distcp
> ----------------------------------------------
>
> Key: GOBBLIN-2159
> URL: https://issues.apache.org/jira/browse/GOBBLIN-2159
> Project: Apache Gobblin
> Issue Type: Task
> Reporter: Vivek Rai
> Priority: Major
> Time Spent: 10h 50m
> Remaining Estimate: 0h
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
