cloud-fan commented on code in PR #57962: URL: https://github.com/apache/spark/pull/57962#discussion_r3797665197
########## sql/core/src/main/scala/org/apache/spark/sql/catalyst/parser/ParseSqlResult.scala: ########## @@ -0,0 +1,382 @@ +/* + * 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.parser + +import scala.collection.mutable + +import org.json4s._ +import org.json4s.jackson.JsonMethods.{compact, parse => parseJson, render} + +import org.apache.spark.{ErrorMessageFormat, SparkThrowable, SparkThrowableHelper} +import org.apache.spark.sql.catalyst.analysis._ +import org.apache.spark.sql.catalyst.expressions._ +import org.apache.spark.sql.catalyst.plans.logical._ +import org.apache.spark.sql.catalyst.trees.{CurrentOrigin, Origin, SQLQueryContext} +import org.apache.spark.sql.exceptions.SqlScriptingException +import org.apache.spark.sql.execution.SparkSqlParser +import org.apache.spark.sql.execution.command.{CreateViewCommand, DescribeQueryCommand, ExplainCommand} +import org.apache.spark.sql.execution.datasources.CreateTempViewUsing + +/** + * Parses a SQL statement string and returns a compact JSON description of the + * unresolved plan (parse-only; no catalog resolution). + * + * Uses a stock [[SparkSqlParser]] (ThreadLocal) so statement coverage matches + * the default production parser (EXPLAIN / SET / ADD JAR / temp views / etc.). + * Session-specific [[org.apache.spark.sql.SparkSessionExtensions]] parser + * wrappers are intentionally not applied: `parse_sql` must evaluate on + * executors without a session, so only the stock parser is available under + * distributed eval. + * + * On success the JSON always includes `parse_success`, the statement + * identifier/code (ISO/IEC 9075-2:2023 Table 39), and omits unused optional + * fields (`table_references`, `function_references`, `select_list`, + * `parameter_markers`) when empty. On parse failure it returns + * `parse_success: false` with source location and a nested STANDARD-format + * error object, and does not throw. Only [[ParseException]] / + * [[SqlScriptingException]] are converted to JSON; unexpected / internal + * failures propagate so the function fails. + */ +object ParseSqlResult { + + private val parser: ThreadLocal[SparkSqlParser] = + ThreadLocal.withInitial(() => new SparkSqlParser()) + + /** Parse `sql` and render the JSON result string. */ + def fromSql(sql: String): String = { + try { + // Do not inherit the outer query's origin from the parse_sql expression. + // Errors and parsed nodes must refer to the SQL string passed to this function. + val origin = if (sql.nonEmpty) { + Origin(startIndex = Some(0), stopIndex = Some(sql.length - 1), sqlText = Some(sql)) + } else { + Origin(sqlText = Some(sql)) + } + CurrentOrigin.withOrigin(origin) { + val plan = parser.get().parsePlan(sql) + fromPlan(plan) + } + } catch { + // User-facing parse / scripting failures become JSON; everything else fails. + case e: ParseException => + errorJson(e) + case e: SqlScriptingException => + errorJson(e) + } + } + + /** Build success JSON from an already-parsed unresolved plan. */ + def fromPlan(plan: LogicalPlan): String = { + val classification = SqlStatementCodes.classify(plan) + val fields = mutable.ListBuffer.empty[JField] + fields += "parse_success" -> JBool(true) + fields += "statement_identifier" -> JString(classification.statementIdentifier) + fields += "statement_code" -> JInt(classification.statementCode) + // Omit unused collections / markers so consumers can treat absence as empty. + val refs = collectPlanReferences(plan) + if (refs.tables.nonEmpty) { + fields += "table_references" -> JArray(refs.tables.map(partsToJArray).toList) + } + if (refs.functions.nonEmpty) { + fields += "function_references" -> JArray(refs.functions.map(partsToJArray).toList) + } + val selectList = collectSelectList(plan) + if (selectList.nonEmpty) { + fields += "select_list" -> JArray(selectList.toList) + } + refs.parameterMarkers.foreach(markers => fields += "parameter_markers" -> markers) + compact(render(JObject(fields.toList))) + } + + private def errorJson(e: SparkThrowable with Throwable): String = { + val errorObj = parseJson( + SparkThrowableHelper.getMessage(e, ErrorMessageFormat.STANDARD)).asInstanceOf[JObject] + val origin = e match { + case p: ParseException => Some(p.start) + case s: SqlScriptingException => Some(s.origin) + case _ => None + } + val locationFields = origin.toSeq.flatMap(originFields) + val contextFields = if (errorObj.obj.exists(_._1 == "queryContext")) { + Nil + } else { + origin.toSeq.flatMap(queryContextField) + } + compact(render(JObject( + "parse_success" -> JBool(false), + "error" -> JObject(errorObj.obj ++ contextFields ++ locationFields) + ))) + } + + private def queryContextField(origin: Origin): Option[JField] = origin.context match { + case context: SQLQueryContext if context.isValid => + Some("queryContext" -> JArray(List(JObject( + "objectType" -> JString(context.objectType), + "objectName" -> JString(context.objectName), + "startIndex" -> JInt(context.startIndex + 1), + "stopIndex" -> JInt(context.stopIndex + 1), + "fragment" -> JString(context.fragment) + )))) + case _ => None + } + + private def originFields(origin: Origin): Seq[JField] = Seq( + origin.line.map(line => "line" -> JInt(line)), + origin.startPosition.map(position => "position" -> JInt(position))).flatten + + private def partsToJArray(parts: Seq[String]): JArray = + JArray(parts.map(JString).toList) + + /** + * Walk expressions in all product fields, including wrappers such as column + * definitions that [[LogicalPlan.expressions]] does not descend into. + */ + private def foreachExpressionDeep(plan: LogicalPlan)(f: Expression => Unit): Unit = { + def visit(value: Any): Unit = value match { + case e: Expression => f(e) + case _: LogicalPlan => + case values: Iterable[_] => values.foreach(visit) + case value: Product => value.productIterator.foreach(visit) + case _ => + } + plan.productIterator.foreach(visit) + } + + /** + * Deep plan walk covering tree slots that standard `collect` / + * `collectWithSubqueries` miss: + * - [[UnresolvedWith]] CTE definitions (`innerChildren`, not `children`) + * - [[InsertIntoStatement]].table (non-child plan slot) + * - [[SingleStatement]].parsedPlan root (children expose only nested + * children; visit the root without re-walking those children) + * - [[CompoundBody]].handlers (not in `children`) + * - [[SimpleCaseStatement]].elseBody (not in `children`) + * Nested expression subqueries are still covered by `foreachWithSubqueries`. + */ + private def foreachPlanDeep(plan: LogicalPlan)(f: LogicalPlan => Unit): Unit = { + plan.foreachWithSubqueries { p => + f(p) + visitNonChildSlots(p)(f) + } + } + + /** + * Visit plan slots that are not exposed via `children` / expression + * subqueries. For [[SingleStatement]], only the wrapped root (and its + * subqueries / non-child slots) is visited -- its children were already + * walked by the outer `foreachWithSubqueries` via + * `SingleStatement.children`. + */ + private def visitNonChildSlots(plan: LogicalPlan)(f: LogicalPlan => Unit): Unit = { + plan match { + case w: UnresolvedWith => + w.cteRelations.foreach { case (_, ctePlan, _) => + foreachPlanDeep(ctePlan)(f) + } + case InsertIntoStatement(table, _, _, _, _, _, _, _, _) => + foreachPlanDeep(table)(f) + case s: SingleStatement => + visitSingleStatementRoot(s.parsedPlan)(f) + case c: CompoundBody => + c.handlers.foreach(h => foreachPlanDeep(h)(f)) + case s: SimpleCaseStatement => + s.elseBody.foreach(b => foreachPlanDeep(b)(f)) + case ExplainCommand(logicalPlan, _) => + foreachPlanDeep(logicalPlan)(f) + case DescribeQueryCommand(_, queryPlan) => + foreachPlanDeep(queryPlan)(f) + case _ => + } + } + + /** + * Visit a [[SingleStatement]] wrapped root without re-traversing its + * children (already exposed by `SingleStatement.children`). + */ + private def visitSingleStatementRoot(root: LogicalPlan)(f: LogicalPlan => Unit): Unit = { + f(root) + root.subqueries.foreach(sq => foreachPlanDeep(sq)(f)) + visitNonChildSlots(root)(f) + } + + /** Multipart name from a table/view-shaped plan node, if any. */ + private def tableOrViewParts(plan: LogicalPlan): Option[Seq[String]] = plan match { + case u: UnresolvedRelation => Some(u.multipartIdentifier) + case u: UnresolvedTable => Some(u.multipartIdentifier) + case u: UnresolvedView => Some(u.multipartIdentifier) + case u: UnresolvedTableOrView => Some(u.multipartIdentifier) + case u: UnresolvedIdentifier => Some(u.nameParts) + case _ => None + } + + private def tableIdentifierParts(id: org.apache.spark.sql.catalyst.TableIdentifier): Seq[String] = + id.catalog.toSeq ++ id.database.toSeq :+ id.table + + private final case class PlanReferences( + tables: Seq[Seq[String]], + functions: Seq[Seq[String]], + parameterMarkers: Option[JObject]) + + /** + * Collect multipart table/view identifiers, function names, and parameter + * markers for lineage. CTE definition names and correlation aliases are + * omitted from tables; tables referenced inside CTE bodies are still + * included. Function / variable identifiers are not collected as tables. + * Deduplicates while preserving first-seen order. Uses one CTE-name pass + * then one combined deep walk for tables, functions, and parameters. + */ + private def collectPlanReferences(plan: LogicalPlan): PlanReferences = { + val tables = mutable.LinkedHashSet.empty[Seq[String]] + val functions = mutable.LinkedHashSet.empty[Seq[String]] + val namedParams = mutable.LinkedHashSet.empty[String] + var unnamedCount = 0 + // CTE names are always single-part in the grammar; UnresolvedRelation refs + // to CTEs are likewise single-part, so filtering matches that shape. + val cteNames = mutable.HashSet.empty[String] Review Comment: **Blocking:** Keep CTE-name filtering scoped to the `UnresolvedWith` that defines each alias. For example, `SELECT * FROM real_t WHERE EXISTS (WITH real_t AS (SELECT * FROM inner_base) SELECT * FROM real_t)` should report both `real_t` and `inner_base`, but this statement-wide set suppresses the outer real table too. Please carry the active CTE environment during traversal and add this shadowing case to the lineage tests. -- 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]
