[HIVE-11433] NPE for a multiple inner join query

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.1.0, 1.2.0, 2.0.0
• Fix Version/s: 1.3.0, 2.0.0
• Component/s: Query Processor
• Assignee: Xuefu Zhang
• Reporter: Xuefu Zhang
• Created: 01/Aug/15 06:30
• Updated: 16/Feb/16 23:50
• Resolved: 05/Aug/15 17:24

Description

NullPointException is thrown for query that has multiple (greater than 3) inner joins. Stacktrace for 1.1.0

NullPointerException null
java.lang.NullPointerException
        at org.apache.hadoop.hive.ql.parse.ParseUtils.getIndex(ParseUtils.java:149)
        at org.apache.hadoop.hive.ql.parse.ParseUtils.checkJoinFilterRefersOneAlias(ParseUtils.java:166)
        at org.apache.hadoop.hive.ql.parse.ParseUtils.checkJoinFilterRefersOneAlias(ParseUtils.java:185)
        at org.apache.hadoop.hive.ql.parse.ParseUtils.checkJoinFilterRefersOneAlias(ParseUtils.java:185)
        at org.apache.hadoop.hive.ql.parse.SemanticAnalyzer.mergeJoins(SemanticAnalyzer.java:8257)
        at org.apache.hadoop.hive.ql.parse.SemanticAnalyzer.mergeJoinTree(SemanticAnalyzer.java:8422)
        at org.apache.hadoop.hive.ql.parse.SemanticAnalyzer.genPlan(SemanticAnalyzer.java:9805)
        at org.apache.hadoop.hive.ql.parse.SemanticAnalyzer.genPlan(SemanticAnalyzer.java:9714)
        at org.apache.hadoop.hive.ql.parse.SemanticAnalyzer.genOPTree(SemanticAnalyzer.java:10150)
        at org.apache.hadoop.hive.ql.parse.SemanticAnalyzer.analyzeInternal(SemanticAnalyzer.java:10161)
        at org.apache.hadoop.hive.ql.parse.SemanticAnalyzer.analyzeInternal(SemanticAnalyzer.java:10078)
        at org.apache.hadoop.hive.ql.parse.BaseSemanticAnalyzer.analyze(BaseSemanticAnalyzer.java:222)
        at org.apache.hadoop.hive.ql.Driver.compile(Driver.java:421)
        at org.apache.hadoop.hive.ql.Driver.compile(Driver.java:307)
        at org.apache.hadoop.hive.ql.Driver.compileInternal(Driver.java:1110)
        at org.apache.hadoop.hive.ql.Driver.compileAndRespond(Driver.java:1104)
        at org.apache.hive.service.cli.operation.SQLOperation.prepare(SQLOperation.java:101)
        at org.apache.hive.service.cli.operation.SQLOperation.runInternal(SQLOperation.java:172)
        at org.apache.hive.service.cli.operation.Operation.run(Operation.java:257)
        at org.apache.hive.service.cli.session.HiveSessionImpl.executeStatementInternal(HiveSessionImpl.java:386)
        at org.apache.hive.service.cli.session.HiveSessionImpl.executeStatementAsync(HiveSessionImpl.java:373)
        at org.apache.hive.service.cli.CLIService.executeStatementAsync(CLIService.java:271)
        at org.apache.hive.service.cli.thrift.ThriftCLIService.ExecuteStatement(ThriftCLIService.java:486)
        at org.apache.hive.service.cli.thrift.TCLIService$Processor$ExecuteStatement.getResult(TCLIService.java:1313)
        at org.apache.hive.service.cli.thrift.TCLIService$Processor$ExecuteStatement.getResult(TCLIService.java:1298)
        at org.apache.thrift.ProcessFunction.process(ProcessFunction.java:39)
        at org.apache.thrift.TBaseProcessor.process(TBaseProcessor.java:39)
        at org.apache.hadoop.hive.thrift.HadoopThriftAuthBridge$Server$TUGIAssumingProcessor.process(HadoopThriftAuthBridge.java:692)
        at org.apache.thrift.server.TThreadPoolServer$WorkerProcess.run(TThreadPoolServer.java:285)
        at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
        at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
        at java.lang.Thread.run(Thread.java:745)

.
However, the problem can also be reproduced in latest master branch. Further investigation shows that the following code (in ParseUtils.java) is problematic:

  static int getIndex(String[] list, String elem) {
    for(int i=0; i < list.length; i++) {
      if (list[i].toLowerCase().equals(elem)) {
        return i;
      }
    }
    return -1;
  }

The code assumes that every element in the list is not null, which isn't true because of the following code in SemanticAnalyzer.java (method genJoinTree()):

    if ((right.getToken().getType() == HiveParser.TOK_TABREF)
        || (right.getToken().getType() == HiveParser.TOK_SUBQUERY)
        || (right.getToken().getType() == HiveParser.TOK_PTBLFUNCTION)) {
      String tableName = getUnescapedUnqualifiedTableName((ASTNode) right.getChild(0))
          .toLowerCase();
      String alias = extractJoinAlias(right, tableName);
      String[] rightAliases = new String[1];
      rightAliases[0] = alias;
      joinTree.setRightAliases(rightAliases);
      String[] children = joinTree.getBaseSrc();
      if (children == null) {
        children = new String[2];
      }
      children[1] = alias;
      joinTree.setBaseSrc(children);
      joinTree.setId(qb.getId());
      joinTree.getAliasToOpInfo().put(
          getModifiedAlias(qb, alias), aliasToOpInfo.get(alias));
      // remember rhs table for semijoin
      if (joinTree.getNoSemiJoin() == false) {
        joinTree.addRHSSemijoin(alias);
      }
    } else {

.
Specifically, this code can result a null element as base source:

      if (children == null) {
        children = new String[2];
      }
      children[1] = alias;

This appears to be a regression from earlier release (0.14.1). However, it's unclear which commit caused this.