hequn8128 commented on a change in pull request #9766: [FLINK-14018][python] Add Python building blocks to make sure the basic functionality of Python ScalarFunction could work URL: https://github.com/apache/flink/pull/9766#discussion_r328533791
########## File path: flink-table/flink-table-planner/src/main/scala/org/apache/flink/table/plan/nodes/CommonPythonCalc.scala ########## @@ -0,0 +1,72 @@ +/* + * 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.flink.table.plan.nodes + +import org.apache.calcite.rex.{RexCall, RexInputRef, RexNode} +import org.apache.flink.table.functions.FunctionLanguage +import org.apache.flink.table.functions.python.{PythonFunction, PythonFunctionInfo, SimplePythonFunction} +import org.apache.flink.table.functions.utils.ScalarSqlFunction + +import scala.collection.JavaConversions._ +import scala.collection.mutable + +trait CommonPythonCalc { + + private[flink] def extractPythonScalarFunctionInfos( + rexCalls: Array[RexCall]): (Array[Int], Array[PythonFunctionInfo]) = { + // using LinkedHashMap to keep the insert order + val inputNodes = new mutable.LinkedHashMap[RexNode, Integer]() + val pythonFunctionInfos = rexCalls.map(createPythonScalarFunctionInfo(_, inputNodes)) + + val udfInputOffsets = inputNodes.toArray.sortBy(_._2).map(_._1).map { + case inputRef: RexInputRef => inputRef.getIndex + } + (udfInputOffsets, pythonFunctionInfos) + } + + private[flink] def createPythonScalarFunctionInfo( + rexCall: RexCall, + inputNodes: mutable.Map[RexNode, Integer]): PythonFunctionInfo = rexCall.getOperator match { + case sfc: ScalarSqlFunction if sfc.getScalarFunction.getLanguage == FunctionLanguage.PYTHON => + val inputs = new mutable.ArrayBuffer[AnyRef]() + rexCall.getOperands.foreach { + case pythonRexCall: RexCall if pythonRexCall.getOperator.asInstanceOf[ScalarSqlFunction] + .getScalarFunction.getLanguage == FunctionLanguage.PYTHON => + // Continuous Python UDFs can be chained together + val argPythonInfo = createPythonScalarFunctionInfo(pythonRexCall, inputNodes) + inputs.append(argPythonInfo) + + case argNode: RexNode => Review comment: Should we add exceptions and meaningful exception messages here if the RexNode is a RexLiteral? We don't support UDFs with literals now. For example, we can do the check and throw the exception in the DataStreamPythonCalcRule, thus we can check it through the plan test. Or can we throw the exception earlier? ---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
