Skip to content

Spark: Change Delete granularity to file for Spark 3.5 #11478

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
amogh-jahagirdar merged 1 commit into from
Jan 3, 2025
Merged

Spark: Change Delete granularity to file for Spark 3.5 #11478

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
10 changes: 6 additions & 4 deletions ...k/v3.5/spark-extensions/src/test/java/org/apache/iceberg/spark/extensions/TestDelete.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -63,6 +63,7 @@
import org.apache.iceberg.Table;
import org.apache.iceberg.TableProperties;
import org.apache.iceberg.data.GenericRecord;
import org.apache.iceberg.deletes.DeleteGranularity;
import org.apache.iceberg.exceptions.ValidationException;
import org.apache.iceberg.relocated.com.google.common.collect.ImmutableList;
import org.apache.iceberg.relocated.com.google.common.collect.ImmutableMap;
Expand Down Expand Up @@ -151,12 +152,15 @@ public void testCoalesceDelete() throws Exception {

// set the open file cost large enough to produce a separate scan task per file
// use range distribution to trigger a shuffle
// set partitioned scoped deletes so that 1 delete file is written as part of the output task
Map<String, String> tableProps =
ImmutableMap.of(
SPLIT_OPEN_FILE_COST,
String.valueOf(Integer.MAX_VALUE),
DELETE_DISTRIBUTION_MODE,
DistributionMode.RANGE.modeName());
DistributionMode.RANGE.modeName(),
TableProperties.DELETE_GRANULARITY,
DeleteGranularity.PARTITION.toString());
sql("ALTER TABLE %s SET TBLPROPERTIES (%s)", tableName, tablePropsAsString(tableProps));

createBranchIfNeeded();
Expand Down Expand Up @@ -1306,10 +1310,8 @@ public void testDeleteWithMultipleSpecs() {
Snapshot currentSnapshot = SnapshotUtil.latestSnapshot(table, branch);
if (mode(table) == COPY_ON_WRITE) {
validateCopyOnWrite(currentSnapshot, "3", "4", "1");
} else if (mode(table) == MERGE_ON_READ && formatVersion >= 3) {
validateMergeOnRead(currentSnapshot, "3", "4", null);
} else {
validateMergeOnRead(currentSnapshot, "3", "3", null);
validateMergeOnRead(currentSnapshot, "3", "4", null);
}

assertEquals(
Expand Down
10 changes: 8 additions & 2 deletions spark/v3.5/spark-extensions/src/test/java/org/apache/iceberg/spark/extensions/TestMerge.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -19,6 +19,7 @@
package org.apache.iceberg.spark.extensions;

import static org.apache.iceberg.RowLevelOperationMode.COPY_ON_WRITE;
import static org.apache.iceberg.RowLevelOperationMode.MERGE_ON_READ;
import static org.apache.iceberg.TableProperties.COMMIT_MAX_RETRY_WAIT_MS;
import static org.apache.iceberg.TableProperties.COMMIT_MIN_RETRY_WAIT_MS;
import static org.apache.iceberg.TableProperties.MERGE_DISTRIBUTION_MODE;
Expand Down Expand Up @@ -56,6 +57,7 @@
import org.apache.iceberg.Table;
import org.apache.iceberg.TableProperties;
import org.apache.iceberg.data.GenericRecord;
import org.apache.iceberg.deletes.DeleteGranularity;
import org.apache.iceberg.exceptions.ValidationException;
import org.apache.iceberg.relocated.com.google.common.collect.ImmutableList;
import org.apache.iceberg.relocated.com.google.common.collect.ImmutableMap;
Expand Down Expand Up @@ -233,7 +235,6 @@ public void testMergeWithVectorizedReads() {

@TestTemplate
public void testCoalesceMerge() {
assumeThat(formatVersion).isLessThan(3);
Copy link
Contributor Author

@amogh-jahagirdar amogh-jahagirdar Dec 12, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Addressing a miss from my previous PR for writing DVs in V3, instead of skipping this coalesce case when the format version is less than 3, we should be asserting the expected DVs (the else if case below)

createAndInitTable("id INT, salary INT, dep STRING");

String[] records = new String[100];
Expand All @@ -252,7 +253,9 @@ public void testCoalesceMerge() {
SPLIT_OPEN_FILE_COST,
String.valueOf(Integer.MAX_VALUE),
MERGE_DISTRIBUTION_MODE,
DistributionMode.NONE.modeName());
DistributionMode.NONE.modeName(),
TableProperties.DELETE_GRANULARITY,
DeleteGranularity.PARTITION.toString());
sql("ALTER TABLE %s SET TBLPROPERTIES (%s)", tableName, tablePropsAsString(tableProps));

createBranchIfNeeded();
Expand Down Expand Up @@ -295,6 +298,9 @@ public void testCoalesceMerge() {
// AQE detects that all shuffle blocks are small and processes them in 1 task
// otherwise, there would be 200 tasks writing to the table
validateProperty(currentSnapshot, SnapshotSummary.ADDED_FILES_PROP, "1");
} else if (mode(table) == MERGE_ON_READ && formatVersion >= 3) {
validateProperty(currentSnapshot, SnapshotSummary.ADDED_DELETE_FILES_PROP, "4");
validateProperty(currentSnapshot, SnapshotSummary.ADDED_DVS_PROP, "4");
} else {
// MoR MERGE would perform a join on `id`
// every task has data for each of 200 reducers
Expand Down
2 changes: 1 addition & 1 deletion ...est/java/org/apache/iceberg/spark/extensions/TestRewritePositionDeleteFilesProcedure.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -49,7 +49,7 @@ private void createTable(boolean partitioned) throws Exception {
String partitionStmt = partitioned ? "PARTITIONED BY (id)" : "";
sql(
"CREATE TABLE %s (id bigint, data string) USING iceberg %s TBLPROPERTIES"
+ "('format-version'='2', 'write.delete.mode'='merge-on-read')",
+ "('format-version'='2', 'write.delete.mode'='merge-on-read', 'write.delete.granularity'='partition')",
tableName, partitionStmt);

List<SimpleRecord> records =
Expand Down
10 changes: 6 additions & 4 deletions ...k/v3.5/spark-extensions/src/test/java/org/apache/iceberg/spark/extensions/TestUpdate.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -58,6 +58,7 @@
import org.apache.iceberg.Table;
import org.apache.iceberg.TableProperties;
import org.apache.iceberg.data.GenericRecord;
import org.apache.iceberg.deletes.DeleteGranularity;
import org.apache.iceberg.exceptions.ValidationException;
import org.apache.iceberg.relocated.com.google.common.collect.ImmutableList;
import org.apache.iceberg.relocated.com.google.common.collect.ImmutableMap;
Expand Down Expand Up @@ -131,12 +132,15 @@ public void testCoalesceUpdate() {

// set the open file cost large enough to produce a separate scan task per file
// use range distribution to trigger a shuffle
// set partitioned scoped deletes so that 1 delete file is written as part of the output task
Map<String, String> tableProps =
ImmutableMap.of(
SPLIT_OPEN_FILE_COST,
String.valueOf(Integer.MAX_VALUE),
UPDATE_DISTRIBUTION_MODE,
DistributionMode.RANGE.modeName());
DistributionMode.RANGE.modeName(),
TableProperties.DELETE_GRANULARITY,
DeleteGranularity.PARTITION.toString());
sql("ALTER TABLE %s SET TBLPROPERTIES (%s)", tableName, tablePropsAsString(tableProps));

createBranchIfNeeded();
Expand Down Expand Up @@ -440,10 +444,8 @@ public void testUpdateWithoutCondition() {
validateProperty(currentSnapshot, CHANGED_PARTITION_COUNT_PROP, "2");
validateProperty(currentSnapshot, DELETED_FILES_PROP, "3");
validateProperty(currentSnapshot, ADDED_FILES_PROP, ImmutableSet.of("2", "3"));
} else if (mode(table) == MERGE_ON_READ && formatVersion >= 3) {
validateMergeOnRead(currentSnapshot, "2", "3", "2");
} else {
validateMergeOnRead(currentSnapshot, "2", "2", "2");
validateMergeOnRead(currentSnapshot, "2", "3", "2");
}

assertEquals(
Expand Down
14 changes: 6 additions & 8 deletions spark/v3.5/spark/src/main/java/org/apache/iceberg/spark/SparkWriteConf.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -714,14 +714,12 @@ private double shuffleCompressionRatio(FileFormat outputFileFormat, String outpu
}

public DeleteGranularity deleteGranularity() {
String valueAsString =
confParser
.stringConf()
.option(SparkWriteOptions.DELETE_GRANULARITY)
.tableProperty(TableProperties.DELETE_GRANULARITY)
.defaultValue(TableProperties.DELETE_GRANULARITY_DEFAULT)
.parse();
return DeleteGranularity.fromString(valueAsString);
return confParser
.enumConf(DeleteGranularity::fromString)
.option(SparkWriteOptions.DELETE_GRANULARITY)
.tableProperty(TableProperties.DELETE_GRANULARITY)
.defaultValue(DeleteGranularity.FILE)
.parse();
}

public boolean useDVs() {
Expand Down
6 changes: 3 additions & 3 deletions spark/v3.5/spark/src/test/java/org/apache/iceberg/spark/TestSparkWriteConf.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -142,7 +142,7 @@ public void testDeleteGranularityDefault() {
SparkWriteConf writeConf = new SparkWriteConf(spark, table, ImmutableMap.of());

DeleteGranularity value = writeConf.deleteGranularity();
assertThat(value).isEqualTo(DeleteGranularity.PARTITION);
assertThat(value).isEqualTo(DeleteGranularity.FILE);
}

@TestTemplate
Expand All @@ -151,13 +151,13 @@ public void testDeleteGranularityTableProperty() {

table
.updateProperties()
.set(TableProperties.DELETE_GRANULARITY, DeleteGranularity.FILE.toString())
.set(TableProperties.DELETE_GRANULARITY, DeleteGranularity.PARTITION.toString())
.commit();

SparkWriteConf writeConf = new SparkWriteConf(spark, table, ImmutableMap.of());

DeleteGranularity value = writeConf.deleteGranularity();
assertThat(value).isEqualTo(DeleteGranularity.FILE);
assertThat(value).isEqualTo(DeleteGranularity.PARTITION);
}

@TestTemplate
Expand Down
4 changes: 3 additions & 1 deletion .../src/test/java/org/apache/iceberg/spark/actions/TestRewritePositionDeleteFilesAction.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -757,7 +757,9 @@ private Map<String, String> tableProperties() {
TableProperties.FORMAT_VERSION,
"2",
TableProperties.DEFAULT_FILE_FORMAT,
format.toString());
format.toString(),
TableProperties.DELETE_GRANULARITY,
DeleteGranularity.PARTITION.toString());
Comment on lines +761 to +762
Copy link
Contributor Author

@amogh-jahagirdar amogh-jahagirdar Nov 11, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This is just the default setup for a table in this test, we do test file scoped deletes as well for this action in testFileGranularity in this test class.

}

private void writeRecords(Table table, int files, int numRecords) {
Expand Down