deniskuzZ commented on code in PR #6039:
URL: https://github.com/apache/hive/pull/6039#discussion_r2293084110
##########
iceberg/iceberg-handler/src/main/java/org/apache/iceberg/mr/hive/compaction/IcebergTableOptimizer.java:
##########
@@ -85,52 +85,53 @@ public IcebergTableOptimizer(HiveConf conf, TxnStore
txnHandler, MetadataCache m
*/
@Override
public Set<CompactionInfo> findPotentialCompactions(long lastChecked,
ShowCompactResponse currentCompactions,
- Set<String> skipDBs, Set<String> skipTables) {
+ Set<String> skipDBs,
Set<String> skipTables) {
Set<CompactionInfo> compactionTargets = Sets.newHashSet();
- getTableNames().stream()
- .filter(table -> !skipDBs.contains(table.getDb()))
- .filter(table -> !skipTables.contains(table.getNotEmptyDbTable()))
- .map(table -> {
- org.apache.hadoop.hive.ql.metadata.Table hiveTable =
getHiveTable(table.getDb(), table.getTable());
- org.apache.iceberg.Table icebergTable =
IcebergTableUtil.getTable(conf, hiveTable.getTTable());
- return Pair.of(hiveTable, icebergTable);
- })
- .filter(t -> hasNewCommits(t.getRight(),
- snapshotTimeMilCache.get(t.getLeft().getFullyQualifiedName())))
- .forEach(t -> {
- String qualifiedTableName = t.getLeft().getFullyQualifiedName();
- org.apache.hadoop.hive.ql.metadata.Table hiveTable = t.getLeft();
- org.apache.iceberg.Table icebergTable = t.getRight();
-
- if (icebergTable.spec().isPartitioned()) {
- List<org.apache.hadoop.hive.ql.metadata.Partition> partitions =
findModifiedPartitions(hiveTable,
- icebergTable, snapshotTimeMilCache.get(qualifiedTableName),
true);
-
- partitions.forEach(partition ->
addCompactionTargetIfEligible(hiveTable.getTTable(), icebergTable,
- partition.getName(), compactionTargets, currentCompactions,
skipDBs, skipTables));
-
- if (IcebergTableUtil.hasUndergonePartitionEvolution(icebergTable)
&& !findModifiedPartitions(hiveTable,
- icebergTable, snapshotTimeMilCache.get(qualifiedTableName),
false).isEmpty()) {
- addCompactionTargetIfEligible(hiveTable.getTTable(),
icebergTable,
- null, compactionTargets, currentCompactions, skipDBs,
skipTables);
- }
- } else {
- addCompactionTargetIfEligible(hiveTable.getTTable(), icebergTable,
null, compactionTargets,
- currentCompactions, skipDBs, skipTables);
- }
-
- snapshotTimeMilCache.put(qualifiedTableName,
icebergTable.currentSnapshot().timestampMillis());
- });
+ Iterable<Table> tables = getTables();
Review Comment:
IcebergTableUtil.getTable would be making 1 HMS call per table. Would be
nice if we can optimize by caching the data (i.e. use
HiveMetaStoreClientWithLocalCache).
Catalogs.loadTable -> HiveTableOperations.doRefresh ->
metaClients.run(client -> client.getTable(database, tableName));
##########
iceberg/iceberg-handler/src/main/java/org/apache/iceberg/mr/hive/compaction/IcebergTableOptimizer.java:
##########
@@ -85,52 +85,53 @@ public IcebergTableOptimizer(HiveConf conf, TxnStore
txnHandler, MetadataCache m
*/
@Override
public Set<CompactionInfo> findPotentialCompactions(long lastChecked,
ShowCompactResponse currentCompactions,
- Set<String> skipDBs, Set<String> skipTables) {
+ Set<String> skipDBs,
Set<String> skipTables) {
Set<CompactionInfo> compactionTargets = Sets.newHashSet();
- getTableNames().stream()
- .filter(table -> !skipDBs.contains(table.getDb()))
- .filter(table -> !skipTables.contains(table.getNotEmptyDbTable()))
- .map(table -> {
- org.apache.hadoop.hive.ql.metadata.Table hiveTable =
getHiveTable(table.getDb(), table.getTable());
- org.apache.iceberg.Table icebergTable =
IcebergTableUtil.getTable(conf, hiveTable.getTTable());
- return Pair.of(hiveTable, icebergTable);
- })
- .filter(t -> hasNewCommits(t.getRight(),
- snapshotTimeMilCache.get(t.getLeft().getFullyQualifiedName())))
- .forEach(t -> {
- String qualifiedTableName = t.getLeft().getFullyQualifiedName();
- org.apache.hadoop.hive.ql.metadata.Table hiveTable = t.getLeft();
- org.apache.iceberg.Table icebergTable = t.getRight();
-
- if (icebergTable.spec().isPartitioned()) {
- List<org.apache.hadoop.hive.ql.metadata.Partition> partitions =
findModifiedPartitions(hiveTable,
- icebergTable, snapshotTimeMilCache.get(qualifiedTableName),
true);
-
- partitions.forEach(partition ->
addCompactionTargetIfEligible(hiveTable.getTTable(), icebergTable,
- partition.getName(), compactionTargets, currentCompactions,
skipDBs, skipTables));
-
- if (IcebergTableUtil.hasUndergonePartitionEvolution(icebergTable)
&& !findModifiedPartitions(hiveTable,
- icebergTable, snapshotTimeMilCache.get(qualifiedTableName),
false).isEmpty()) {
- addCompactionTargetIfEligible(hiveTable.getTTable(),
icebergTable,
- null, compactionTargets, currentCompactions, skipDBs,
skipTables);
- }
- } else {
- addCompactionTargetIfEligible(hiveTable.getTTable(), icebergTable,
null, compactionTargets,
- currentCompactions, skipDBs, skipTables);
- }
-
- snapshotTimeMilCache.put(qualifiedTableName,
icebergTable.currentSnapshot().timestampMillis());
- });
+ Iterable<Table> tables = getTables();
Review Comment:
IcebergTableUtil.getTable would be making 1 HMS call per table. Would be
nice if we can optimize by caching the data (i.e. use
`HiveMetaStoreClientWithLocalCache`).
Catalogs.loadTable -> HiveTableOperations.doRefresh ->
metaClients.run(client -> client.getTable(database, tableName));
--
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]
