Akshat-Jain commented on code in PR #16682:
URL: https://github.com/apache/druid/pull/16682#discussion_r1675634450
##########
extensions-core/datasketches/src/main/java/org/apache/druid/query/aggregation/datasketches/hll/sql/HllSketchApproxCountDistinctSqlAggregator.java:
##########
@@ -21,28 +21,68 @@
import org.apache.calcite.sql.SqlAggFunction;
import org.apache.calcite.sql.SqlFunctionCategory;
+import org.apache.calcite.sql.type.CastedLiteralOperandTypeCheckers;
import org.apache.calcite.sql.type.InferTypes;
+import org.apache.calcite.sql.type.OperandTypes;
+import org.apache.calcite.sql.type.SqlSingleOperandTypeChecker;
import org.apache.calcite.sql.type.SqlTypeFamily;
import org.apache.calcite.sql.type.SqlTypeName;
import org.apache.druid.java.util.common.StringEncoding;
+import org.apache.druid.java.util.common.StringUtils;
import org.apache.druid.query.aggregation.AggregatorFactory;
+import
org.apache.druid.query.aggregation.datasketches.hll.HllSketchBuildAggregatorFactory;
+import
org.apache.druid.query.aggregation.datasketches.hll.HllSketchMergeAggregatorFactory;
import
org.apache.druid.query.aggregation.post.FinalizingFieldAccessPostAggregator;
import org.apache.druid.sql.calcite.aggregation.Aggregation;
import org.apache.druid.sql.calcite.aggregation.SqlAggregator;
import org.apache.druid.sql.calcite.expression.OperatorConversions;
+import org.apache.druid.sql.calcite.table.RowSignatures;
import java.util.Collections;
+/**
+ * Approximate count distinct aggregator using HLL sketches.
+ * Supported column types: String, Numeric, HLLSketchMerge, HLLSketchBuild.
+ */
public class HllSketchApproxCountDistinctSqlAggregator extends
HllSketchBaseSqlAggregator implements SqlAggregator
{
public static final String NAME = "APPROX_COUNT_DISTINCT_DS_HLL";
+
+ private static final SqlSingleOperandTypeChecker
AGGREGATED_COLUMN_TYPE_CHECKER = OperandTypes.or(
+ OperandTypes.STRING,
+ OperandTypes.NUMERIC,
+ RowSignatures.complexTypeChecker(HllSketchMergeAggregatorFactory.TYPE),
+ RowSignatures.complexTypeChecker(HllSketchBuildAggregatorFactory.TYPE)
+ );
+
private static final SqlAggFunction FUNCTION_INSTANCE =
OperatorConversions.aggregatorBuilder(NAME)
- .operandNames("column", "lgK", "tgtHllType")
- .operandTypes(SqlTypeFamily.ANY,
SqlTypeFamily.NUMERIC, SqlTypeFamily.STRING)
.operandTypeInference(InferTypes.VARCHAR_1024)
- .requiredOperandCount(1)
- .literalOperands(1, 2)
+ .operandTypeChecker(
+ OperandTypes.or(
+ // APPROX_COUNT_DISTINCT_DS_HLL(column)
+
OperandTypes.and(AGGREGATED_COLUMN_TYPE_CHECKER,
OperandTypes.family(SqlTypeFamily.ANY)),
Review Comment:
@kgyrtkirk I also felt that it's redundant, but it was done this way
throughout the codebase, so I ended up following the pattern. Example, in
`CompressedBigDecimalSqlAggregatorBase.java`, we have the following:
```java
OperandTypes.and(
OperandTypes.sequence(
"'" + name + "(column, size, scale)'",
OperandTypes.ANY,
CastedLiteralOperandTypeCheckers.POSITIVE_INTEGER_LITERAL,
CastedLiteralOperandTypeCheckers.POSITIVE_INTEGER_LITERAL
),
OperandTypes.family(SqlTypeFamily.ANY,
SqlTypeFamily.EXACT_NUMERIC, SqlTypeFamily.EXACT_NUMERIC)
)
```
Why is the above not written as the following?
```java
OperandTypes.sequence(
"'" + name + "(column, size, scale)'",
OperandTypes.ANY,
CastedLiteralOperandTypeCheckers.POSITIVE_INTEGER_LITERAL,
CastedLiteralOperandTypeCheckers.POSITIVE_INTEGER_LITERAL
)
```
--
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]
