Skip to content

API, Spark: Fix aggregation pushdown on struct fields #9176

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
merged 3 commits into from
Jan 31, 2024
Merged

API, Spark: Fix aggregation pushdown on struct fields #9176

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
a71727c
Fix agg pushodwn on struct
amogh-jahagirdar Nov 27, 2023
fa5cfe6
Fix ValueAggregate for nested values. (#254)
rdblue Jan 30, 2024
d9b1230
Cleanup tests
amogh-jahagirdar Jan 31, 2024
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
6 changes: 5 additions & 1 deletion api/src/main/java/org/apache/iceberg/expressions/ValueAggregate.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -60,7 +60,11 @@ public int size() {
@Override
@SuppressWarnings("unchecked")
public <T> T get(int pos, Class<T> javaClass) {
return (T) value;
if (javaClass.isAssignableFrom(StructLike.class)) {
return (T) this;
} else {
return (T) value;
}
}

@Override
Expand Down
122 changes: 122 additions & 0 deletions spark/v3.5/spark/src/test/java/org/apache/iceberg/spark/sql/TestAggregatePushDown.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -23,6 +23,7 @@
import java.math.BigDecimal;
import java.sql.Date;
import java.sql.Timestamp;
import java.util.Arrays;
import java.util.List;
import java.util.Locale;
import org.apache.iceberg.CatalogUtil;
Expand All @@ -36,6 +37,7 @@
import org.apache.iceberg.spark.CatalogTestBase;
import org.apache.iceberg.spark.TestBase;
import org.apache.spark.sql.SparkSession;
import org.assertj.core.api.Assertions;
import org.junit.jupiter.api.AfterEach;
import org.junit.jupiter.api.BeforeAll;
import org.junit.jupiter.api.TestTemplate;
Expand Down Expand Up @@ -478,6 +480,126 @@ public void testAggregateWithComplexType() {
.isFalse();
}

@TestTemplate
public void testAggregationPushdownStructInteger() {
sql("CREATE TABLE %s (id BIGINT, struct_with_int STRUCT<c1:BIGINT>) USING iceberg", tableName);
sql("INSERT INTO TABLE %s VALUES (1, named_struct(\"c1\", NULL))", tableName);
sql("INSERT INTO TABLE %s VALUES (2, named_struct(\"c1\", 2))", tableName);
sql("INSERT INTO TABLE %s VALUES (3, named_struct(\"c1\", 3))", tableName);

String query = "SELECT COUNT(%s), MAX(%s), MIN(%s) FROM %s";
String aggField = "struct_with_int.c1";
assertAggregates(sql(query, aggField, aggField, aggField, tableName), 2L, 3L, 2L);
assertExplainContains(
sql("EXPLAIN " + query, aggField, aggField, aggField, tableName),
"count(struct_with_int.c1)",
"max(struct_with_int.c1)",
"min(struct_with_int.c1)");
}

@TestTemplate
public void testAggregationPushdownNestedStruct() {
sql(
"CREATE TABLE %s (id BIGINT, struct_with_int STRUCT<c1:STRUCT<c2:STRUCT<c3:STRUCT<c4:BIGINT>>>>) USING iceberg",
tableName);
sql(
"INSERT INTO TABLE %s VALUES (1, named_struct(\"c1\", named_struct(\"c2\", named_struct(\"c3\", named_struct(\"c4\", NULL)))))",
tableName);
sql(
"INSERT INTO TABLE %s VALUES (2, named_struct(\"c1\", named_struct(\"c2\", named_struct(\"c3\", named_struct(\"c4\", 2)))))",
tableName);
sql(
"INSERT INTO TABLE %s VALUES (3, named_struct(\"c1\", named_struct(\"c2\", named_struct(\"c3\", named_struct(\"c4\", 3)))))",
tableName);

String query = "SELECT COUNT(%s), MAX(%s), MIN(%s) FROM %s";
String aggField = "struct_with_int.c1.c2.c3.c4";

assertAggregates(sql(query, aggField, aggField, aggField, tableName), 2L, 3L, 2L);

assertExplainContains(
sql("EXPLAIN " + query, aggField, aggField, aggField, tableName),
"count(struct_with_int.c1.c2.c3.c4)",
"max(struct_with_int.c1.c2.c3.c4)",
"min(struct_with_int.c1.c2.c3.c4)");
}

@TestTemplate
public void testAggregationPushdownStructTimestamp() {
sql(
"CREATE TABLE %s (id BIGINT, struct_with_ts STRUCT<c1:TIMESTAMP>) USING iceberg",
tableName);
sql("INSERT INTO TABLE %s VALUES (1, named_struct(\"c1\", NULL))", tableName);
sql(
"INSERT INTO TABLE %s VALUES (2, named_struct(\"c1\", timestamp('2023-01-30T22:22:22Z')))",
tableName);
sql(
"INSERT INTO TABLE %s VALUES (3, named_struct(\"c1\", timestamp('2023-01-30T22:23:23Z')))",
tableName);

String query = "SELECT COUNT(%s), MAX(%s), MIN(%s) FROM %s";
String aggField = "struct_with_ts.c1";

assertAggregates(
sql(query, aggField, aggField, aggField, tableName),
2L,
new Timestamp(1675117403000L),
new Timestamp(1675117342000L));

assertExplainContains(
sql("EXPLAIN " + query, aggField, aggField, aggField, tableName),
"count(struct_with_ts.c1)",
"max(struct_with_ts.c1)",
"min(struct_with_ts.c1)");
}

@TestTemplate
public void testAggregationPushdownOnBucketedColumn() {
sql(
"CREATE TABLE %s (id BIGINT, struct_with_int STRUCT<c1:INT>) USING iceberg PARTITIONED BY (bucket(8, id))",
tableName);

sql("INSERT INTO TABLE %s VALUES (1, named_struct(\"c1\", NULL))", tableName);
sql("INSERT INTO TABLE %s VALUES (null, named_struct(\"c1\", 2))", tableName);
sql("INSERT INTO TABLE %s VALUES (2, named_struct(\"c1\", 3))", tableName);

String query = "SELECT COUNT(%s), MAX(%s), MIN(%s) FROM %s";
String aggField = "id";
assertAggregates(sql(query, aggField, aggField, aggField, tableName), 2L, 2L, 1L);
assertExplainContains(
sql("EXPLAIN " + query, aggField, aggField, aggField, tableName),
"count(id)",
"max(id)",
"min(id)");
}

private void assertAggregates(
List<Object[]> actual, Object expectedCount, Object expectedMax, Object expectedMin) {
Object actualCount = actual.get(0)[0];
Object actualMax = actual.get(0)[1];
Object actualMin = actual.get(0)[2];

Assertions.assertThat(actualCount)
.as("Expected and actual count should equal")
.isEqualTo(expectedCount);
Assertions.assertThat(actualMax)
.as("Expected and actual max should equal")
.isEqualTo(expectedMax);
Assertions.assertThat(actualMin)
.as("Expected and actual min should equal")
.isEqualTo(expectedMin);
}

private void assertExplainContains(List<Object[]> explain, String... expectedFragments) {
String explainString = explain.get(0)[0].toString().toLowerCase(Locale.ROOT);
Arrays.stream(expectedFragments)
.forEach(
fragment ->
Assertions.assertThat(explainString.contains(fragment))
Copy link
Contributor

Choose a reason for hiding this comment

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

this should be Assertions.assertThat(explainString).as(...).contains(fragment). We typically try to avoid usage of isTrue() / isFalse() on assertions like these because they don't provide any contextual insight when an assertion fails.
On the other hand, using assertThat(explainString).as(...).contains(fragment) will always show the content of explainString and fragment in case the assertion fails.
Also the .as() typically needs to be specified before the final assertion and will be ignored otherwise.

.isTrue()
.as("Expected to find plan fragment in explain plan"));
}

@TestTemplate
public void testAggregatePushDownInDeleteCopyOnWrite() {
sql("CREATE TABLE %s (id LONG, data INT) USING iceberg", tableName);
Expand Down