Skip to content
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

Jump to bottom

Deduce type from the aggregators when materializing subquery results #16703

Merged
merged 3 commits into from
Jul 23, 2024
Merged

Deduce type from the aggregators when materializing subquery results #16703

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
942d7f3
init
LakshSingla Jul 8, 2024
7668732
review
LakshSingla Jul 17, 2024
c185e8a
review 2
LakshSingla Jul 22, 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 processing/src/main/java/org/apache/druid/query/groupby/GroupByQueryQueryToolChest.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -811,7 +811,11 @@ public Optional<Sequence<FrameSignaturePair>> resultsAsFrames(
boolean useNestedForUnknownTypes
)
{
RowSignature rowSignature = resultArraySignature(query);
RowSignature rowSignature = query.getResultRowSignature(
query.context().isFinalize(true)
? RowSignature.Finalization.YES
: RowSignature.Finalization.NO
);
Comment on lines +814 to +818
Copy link
Member

@kgyrtkirk kgyrtkirk Jul 15, 2024 •
edited
Loading

Choose a reason for hiding this comment

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

I believe this logic supposed to be inside query.getResultRowSignature; query already knows context() ; why should we tell it from the outside the value of Finalization?

doesn't that work?

Copy link
Contributor Author

Choose a reason for hiding this comment

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

It should work, but I am scared to make that change given that it will affect everything from native and SQL queries. Lemme try making the change and see if there are any failing tests.

Copy link
Contributor Author

Choose a reason for hiding this comment

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

I realised that it shouldn't work. For example - look at GroupByPreShuffleFrameProcessor and GroupByPostShuffleFrameProcessor. The same query requires different finalization modes, since one partially aggregates and we need to intermediate type while the other completely aggregates and finalizes. This information isn't fully captured by the query and needs someone from the outside to tell which finalization mode to use. Therefore we can't trustily determine based on the query context.

RowSignature modifiedRowSignature = useNestedForUnknownTypes
? FrameWriterUtils.replaceUnknownTypesWithNestedColumns(rowSignature)
: rowSignature;
Expand Down
27 changes: 18 additions & 9 deletions ...essing/src/main/java/org/apache/druid/query/timeseries/TimeseriesQueryQueryToolChest.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -427,14 +427,7 @@ public Function<Result<TimeseriesResultValue>, Result<TimeseriesResultValue>> ma
@Override
public RowSignature resultArraySignature(TimeseriesQuery query)
{
RowSignature.Builder rowSignatureBuilder = RowSignature.builder();
rowSignatureBuilder.addTimeColumn();
if (StringUtils.isNotEmpty(query.getTimestampResultField())) {
rowSignatureBuilder.add(query.getTimestampResultField(), ColumnType.LONG);
}
rowSignatureBuilder.addAggregators(query.getAggregatorSpecs(), RowSignature.Finalization.UNKNOWN);
rowSignatureBuilder.addPostAggregators(query.getPostAggregatorSpecs());
return rowSignatureBuilder.build();
return resultSignature(query, RowSignature.Finalization.UNKNOWN);
}

@Override
Expand Down Expand Up @@ -474,7 +467,11 @@ public Optional<Sequence<FrameSignaturePair>> resultsAsFrames(
boolean useNestedForUnknownTypes
)
{
final RowSignature rowSignature = resultArraySignature(query);
final RowSignature rowSignature =
resultSignature(
query,
query.context().isFinalize(true) ? RowSignature.Finalization.YES : RowSignature.Finalization.NO
);
final Pair<Cursor, Closeable> cursorAndCloseable = IterableRowsCursorHelper.getCursorFromSequence(
resultsAsArrays(query, resultSequence),
rowSignature
Expand Down Expand Up @@ -535,4 +532,16 @@ private Function<Result<TimeseriesResultValue>, Result<TimeseriesResultValue>> m
);
};
}

private RowSignature resultSignature(final TimeseriesQuery query, final RowSignature.Finalization finalization)
Copy link
Member

Choose a reason for hiding this comment

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

can this method be moved to be: TimeseriesQuery#getResultSignature (like for GroupByQuery )
or TimeseriesQuery#getRowSignature (like for ScanQuery ) ?

{
RowSignature.Builder rowSignatureBuilder = RowSignature.builder();
rowSignatureBuilder.addTimeColumn();
if (StringUtils.isNotEmpty(query.getTimestampResultField())) {
rowSignatureBuilder.add(query.getTimestampResultField(), ColumnType.LONG);
}
rowSignatureBuilder.addAggregators(query.getAggregatorSpecs(), finalization);
rowSignatureBuilder.addPostAggregators(query.getPostAggregatorSpecs());
return rowSignatureBuilder.build();
}
}
11 changes: 10 additions & 1 deletion processing/src/main/java/org/apache/druid/query/topn/TopNQueryQueryToolChest.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -558,7 +558,16 @@ public Optional<Sequence<FrameSignaturePair>> resultsAsFrames(
boolean useNestedForUnknownTypes
)
{
final RowSignature rowSignature = resultArraySignature(query);
final RowSignature rowSignature =
Copy link
Member

Choose a reason for hiding this comment

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

similarily to TS: - shouldn't this be TopNQuery#getResultRowSignature ?
please also update resultArraySignature to use that method so we are not duplicating logic

RowSignature.builder()
.addTimeColumn()
.addDimensions(Collections.singletonList(query.getDimensionSpec()))
.addAggregators(
query.getAggregatorSpecs(),
query.context().isFinalize(true) ? RowSignature.Finalization.YES : RowSignature.Finalization.NO
)
.addPostAggregators(query.getPostAggregatorSpecs())
.build();
final Pair<Cursor, Closeable> cursorAndCloseable = IterableRowsCursorHelper.getCursorFromSequence(
resultsAsArrays(query, resultSequence),
rowSignature
Expand Down
216 changes: 216 additions & 0 deletions sql/src/test/java/org/apache/druid/sql/calcite/CalciteSubqueryTest.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -51,10 +51,12 @@
import org.apache.druid.query.aggregation.LongMinAggregatorFactory;
import org.apache.druid.query.aggregation.LongSumAggregatorFactory;
import org.apache.druid.query.aggregation.SingleValueAggregatorFactory;
import org.apache.druid.query.aggregation.firstlast.first.StringFirstAggregatorFactory;
import org.apache.druid.query.aggregation.post.ArithmeticPostAggregator;
import org.apache.druid.query.aggregation.post.FieldAccessPostAggregator;
import org.apache.druid.query.dimension.DefaultDimensionSpec;
import org.apache.druid.query.dimension.ExtractionDimensionSpec;
import org.apache.druid.query.expression.TestExprMacroTable;
import org.apache.druid.query.extraction.SubstringDimExtractionFn;
import org.apache.druid.query.groupby.GroupByQuery;
import org.apache.druid.query.groupby.orderby.DefaultLimitSpec;
Expand Down Expand Up @@ -1388,6 +1390,220 @@
);
}

@MethodSource("constructorFeeder")
@ParameterizedTest(name = "{0}")
public void testGroupBySubqueryWithEarliestAggregator(String testName, Map<String, Object> queryContext)
Dismissed Show dismissed Hide dismissed
{
cannotVectorize();

// Note: EARLIEST aggregator is used because the intermediate type "serializablePair" is different from the finalized type
final List<Object[]> expectedResults;
if (NullHandling.replaceWithDefault()) {
expectedResults = ImmutableList.of(
new Object[]{"1", "", "a", "1"},
new Object[]{"10.1", "b", "", "10.1"},
new Object[]{"10.1", "c", "", "10.1"},
new Object[]{"2", "d", "", "2"},
new Object[]{"abc", "", "", "abc"},
new Object[]{"def", "", "abc", "def"}
);
} else {
expectedResults = ImmutableList.of(
new Object[]{"", "a", "a", ""},
new Object[]{"", "b", "a", ""},
new Object[]{"1", "", "a", "1"},
new Object[]{"10.1", "b", null, "10.1"},
new Object[]{"10.1", "c", null, "10.1"},
new Object[]{"2", "d", "", "2"},
new Object[]{"abc", null, null, "abc"},
new Object[]{"def", null, "abc", "def"}
);
}

testQuery(
"SELECT a.dim1, a.dim3, a.e_dim2, b.dim1 "
+ "FROM ("
+ " SELECT dim1, dim3, EARLIEST(dim2) AS e_dim2 "
+ " FROM foo GROUP BY 1, 2 LIMIT 100"
+ ") a "
+ "INNER JOIN foo b ON a.dim1 = b.dim1",
queryContext,
ImmutableList.of(
newScanQueryBuilder()
.dataSource(
JoinDataSource.create(
new QueryDataSource(
GroupByQuery.builder()
.setDataSource("foo")
.setInterval(querySegmentSpec(Intervals.ETERNITY))
.setGranularity(Granularities.ALL)
.setDimensions(
new DefaultDimensionSpec("dim1", "d0", ColumnType.STRING),
new DefaultDimensionSpec("dim3", "d1", ColumnType.STRING)
)
.addAggregator(new StringFirstAggregatorFactory("a0", "dim2", "__time", 1024))
.setLimitSpec(new DefaultLimitSpec(Collections.emptyList(), 100))
.build()
),
new QueryDataSource(
newScanQueryBuilder()
.dataSource("foo")
.intervals(querySegmentSpec(Intervals.ETERNITY))
.columns("dim1")
.build()
),
"j0.",
"(\"d0\" == \"j0.dim1\")",
JoinType.INNER,
null,
TestExprMacroTable.INSTANCE,
null
)
)
.intervals(querySegmentSpec(Intervals.ETERNITY))
.columns("a0", "d0", "d1", "j0.dim1")
.context(QUERY_CONTEXT_DEFAULT)
.build()
),
expectedResults
);
}

@MethodSource("constructorFeeder")
@ParameterizedTest(name = "{0}")
public void testTopNSubqueryWithEarliestAggregator(String testName, Map<String, Object> queryContext)
Dismissed Show dismissed Hide dismissed
{
final List<Object[]> expectedResults;
if (NullHandling.replaceWithDefault()) {
expectedResults = ImmutableList.of(
new Object[]{"1", "a", "1"},
new Object[]{"10.1", "", "10.1"},
new Object[]{"2", "", "2"},
new Object[]{"abc", "", "abc"},
new Object[]{"def", "abc", "def"}
);
} else {
expectedResults = ImmutableList.of(
new Object[]{"", "a", ""},
new Object[]{"1", "a", "1"},
new Object[]{"10.1", null, "10.1"},
new Object[]{"2", "", "2"},
new Object[]{"abc", null, "abc"},
new Object[]{"def", "abc", "def"}
);
}

testQuery(
"SELECT a.dim1, a.e_dim2, b.dim1 "
+ "FROM ("
+ " SELECT dim1, EARLIEST(dim2) AS e_dim2 "
+ " FROM foo "
+ " GROUP BY 1 "
+ " LIMIT 100"
+ ") a "
+ "INNER JOIN foo b ON a.dim1 = b.dim1",
queryContext,
ImmutableList.of(
newScanQueryBuilder()
.dataSource(
JoinDataSource.create(
new QueryDataSource(
new TopNQueryBuilder()
.dataSource("foo")
.dimension(new DefaultDimensionSpec("dim1", "d0", ColumnType.STRING))
.metric(new DimensionTopNMetricSpec(null, StringComparators.LEXICOGRAPHIC))
.threshold(100)
.intervals(querySegmentSpec(Intervals.ETERNITY))
.granularity(Granularities.ALL)
.aggregators(
new StringFirstAggregatorFactory("a0", "dim2", "__time", 1024)
)
.build()
),
new QueryDataSource(
newScanQueryBuilder()
.dataSource("foo")
.intervals(querySegmentSpec(Intervals.ETERNITY))
.columns("dim1")
.build()
),
"j0.",
"(\"d0\" == \"j0.dim1\")",
JoinType.INNER,
null,
TestExprMacroTable.INSTANCE,
null
)
)
.intervals(querySegmentSpec(Intervals.ETERNITY))
.columns("a0", "d0", "j0.dim1")
.context(QUERY_CONTEXT_DEFAULT)
.build()
),
expectedResults
);
}

@MethodSource("constructorFeeder")
@ParameterizedTest(name = "{0}")
public void testTimeseriesSubqueryWithEarliestAggregator(String testName, Map<String, Object> queryContext)
Dismissed Show dismissed Hide dismissed
{
testQuery(
"SELECT a.__time, a.e_dim2, b.__time "
+ "FROM ("
+ " SELECT TIME_FLOOR(\"__time\", 'PT24H') as __time, EARLIEST(dim2) AS e_dim2 "
+ " FROM foo "
+ " GROUP BY 1 "
+ ") a "
+ "INNER JOIN foo b ON a.__time = b.__time",
queryContext,
ImmutableList.of(
newScanQueryBuilder()
.dataSource(
JoinDataSource.create(
new QueryDataSource(
Druids.newTimeseriesQueryBuilder()
.dataSource("foo")
.intervals(querySegmentSpec(Intervals.ETERNITY))
.granularity(new PeriodGranularity(
new Period("PT24H"),
null,
DateTimeZone.UTC
))
.aggregators(new StringFirstAggregatorFactory("a0", "dim2", "__time", 1024))
.build()
),
new QueryDataSource(
newScanQueryBuilder()
.dataSource("foo")
.intervals(querySegmentSpec(Intervals.ETERNITY))
.columns("__time")
.build()
),
"j0.",
"(\"d0\" == \"j0.__time\")",
JoinType.INNER,
null,
TestExprMacroTable.INSTANCE,
null
)
)
.intervals(querySegmentSpec(Intervals.ETERNITY))
.columns("a0", "d0", "j0.__time")
.context(QUERY_CONTEXT_DEFAULT)
.build()
),
ImmutableList.of(
new Object[]{946684800000L, "a", 946684800000L},
new Object[]{946771200000L, NullHandling.defaultStringValue(), 946771200000L},
new Object[]{946857600000L, "", 946857600000L},
new Object[]{978307200000L, "a", 978307200000L},
new Object[]{978393600000L, "abc", 978393600000L},
new Object[]{978480000000L, NullHandling.defaultStringValue(), 978480000000L}
)
);
}

public static class SubqueryComponentSupplier extends SqlTestFramework.StandardComponentSupplier
{

Expand Down
Loading