mkaravel commented on code in PR #40615: URL: https://github.com/apache/spark/pull/40615#discussion_r1164399489
########## sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/aggregate/datasketchesAggregates.scala: ########## @@ -0,0 +1,523 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package org.apache.spark.sql.catalyst.expressions.aggregate + +import java.util.Locale + +import org.apache.datasketches.hll.{HllSketch, TgtHllType, Union} +import org.apache.datasketches.memory.WritableMemory + +import org.apache.spark.sql.catalyst.InternalRow +import org.apache.spark.sql.catalyst.analysis.TypeCheckResult +import org.apache.spark.sql.catalyst.analysis.TypeCheckResult.{DataTypeMismatch, TypeCheckSuccess} +import org.apache.spark.sql.catalyst.expressions.{Expression, ExpressionDescription, Literal} +import org.apache.spark.sql.catalyst.trees.{BinaryLike, TernaryLike} +import org.apache.spark.sql.catalyst.util.TypeUtils.{toSQLExpr, toSQLId, toSQLType, toSQLValue} +import org.apache.spark.sql.types.{BinaryType, DataType, IntegerType, LongType, NullType, StringType} +import org.apache.spark.unsafe.types.UTF8String + +/** + * This datasketchesAggregates file is intended to encapsulate all of the + * aggregate functions that utilize Datasketches sketch objects as intermediate + * aggregation buffers. + * + * The HllSketchAggregate sealed trait is meant to be extended by the aggregate + * functions which utilize instances of HllSketch to count uniques. + */ +sealed trait HllSketchAggregate + extends TypedImperativeAggregate[HllSketch] with TernaryLike[Expression] { + + // Hllsketch config - mark as lazy so that they're not evaluated during tree transformation. + + lazy val lgConfigK: Int = second.eval().asInstanceOf[Int] + lazy val tgtHllType: TgtHllType = + TgtHllType.valueOf(third.eval().asInstanceOf[UTF8String].toString.toUpperCase(Locale.ROOT)) + + // Type checking + + override def checkInputDataTypes(): TypeCheckResult = { + (first.dataType, second.dataType, third.dataType) match { + case (_, NullType, _) | (_, _, NullType) => + DataTypeMismatch( + errorSubClass = "UNEXPECTED_NULL", + messageParameters = Map( + "exprName" -> "lgConfigK or tgtHllType" + ) + ) + case (_, IntegerType, StringType) => + if (!second.foldable) { + DataTypeMismatch( + errorSubClass = "NON_FOLDABLE_INPUT", + messageParameters = Map( + "inputName" -> "lgConfigK", + "inputType" -> toSQLType(second.dataType), + "inputExpr" -> toSQLExpr(second) + ) + ) + } else if (lgConfigK <= 0L) { + DataTypeMismatch( + errorSubClass = "VALUE_OUT_OF_RANGE", + messageParameters = Map( + "exprName" -> "lgConfigK", + "valueRange" -> s"[0, positive]", + "currentValue" -> toSQLValue(lgConfigK, IntegerType) + ) + ) + } else if (!third.foldable) { + DataTypeMismatch( + errorSubClass = "NON_FOLDABLE_INPUT", + messageParameters = Map( + "inputName" -> "numBitsExpression", + "inputType" -> toSQLType(third.dataType), + "inputExpr" -> toSQLExpr(third) + ) + ) + } else { + TypeCheckSuccess + } + case _ => + DataTypeMismatch( + errorSubClass = "HLLSKETCH_WRONG_TYPE", + messageParameters = Map( + "functionName" -> toSQLId(prettyName), + "expectedSecond" -> toSQLType(IntegerType), + "expectedThird" -> toSQLType(StringType), + "actual" -> Seq(first.dataType, second.dataType, third.dataType) + .map(toSQLType).mkString(", ") + ) + ) + } + } + + // From here on, these are the shared default implementations for TypedImperativeAggregate + + /** Aggregate functions which utilize HllSketch instances should never return null */ + override def nullable: Boolean = false + + /** + * Instantiate an HllSketch instance using the lgConfigK and tgtHllType params. + * + * @return an HllSketch instance + */ + override def createAggregationBuffer(): HllSketch = { + new HllSketch(lgConfigK, tgtHllType) + } + + /** + * Evaluate the input row and update the HllSketch instance with the row's value. + * The update function only supports a subset of Spark SQL types, and an + * UnsupportedOperationException will be thrown for unsupported types. + * + * @param sketch The HllSketch instance. + * @param input an input row + */ + override def update(sketch: HllSketch, input: InternalRow): HllSketch = { + val v = first.eval(input) + if (v != null) { + first.dataType match { + // Update implemented for a subset of types supported by HllSketch + // Spark SQL doesn't have equivalent types for ByteBuffer or char[] so leave those out + // Leaving out support for Array types, as unique counting these aren't a common use case + // Leaving out support for floating point types (IE DoubleType) due to imprecision + // TODO: implement support for decimal/datetime/interval types + case IntegerType => sketch.update(v.asInstanceOf[Int]) + case LongType => sketch.update(v.asInstanceOf[Long]) + case StringType => sketch.update(v.asInstanceOf[UTF8String].toString) Review Comment: Regarding BINARY values: I think there is a small confusion here: in Spark we have STRING, BINARY, ARRAY, STRUCT, and MAP types. STRING and BINARY are basically arrays of characters/bytes and they are variable size data types. They are not nested, and they are essentially the same (STRINGs have additional semantics as to what should be stored, like UTF8 characters and more). ARRAY, STRUCT, and MAP are nested data types that depend on other data types (the elements of the these nested structures). We cannot really fully define them without knowing the nested types. ARRAY is variable size, STRUCT can be either fixed size or variable size (depending on the type of elements in the struct) and MAP is also variable size. All these from the point of view of types. Then we get to the representation of those types. STRING is mapped to the `UTF8String` class as the underlying "physical" type, and BINARY is mapped to an array of bytes (what you refer to as `Array[Byte]`. BINARY is not treated as an ARRAY in the Spark type system, but rather as a single non-nested type whose "physical" representation is a Scala/Java array of bytes. So in that respect it is not any different from STRING really. -- 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]
