pentaho/pentaho-kettle · error · KettleException

MergeJoin.Exception.UnableToFindSpecifiedStep

MergeJoin.Exception.UnableToFindSpecifiedStep

Error message

MergeJoin.Exception.UnableToFindSpecifiedStep

What it means

MergeJoin.processRow() locates the RowSet for its first (reference) input stream by step name via findInputRowSet(). When it returns null — no such input rowset is registered for this step — it throws this KettleException naming the missing step. The merge join cannot read sorted input from its first hop, so processing aborts.

Solutions

1. Open the Merge Join step dialog and verify both input steps listed exist and are connected with hops.
2. Re-create the hop from the first (reference) source step to the Merge Join step after any rename.
3. Ensure both source steps are enabled and part of the executed transformation.
4. If building the transformation in code, call meta.getStepIOMeta().getInfoStreams() updates and set correct step names before execution.

Example fix

// before: stale step name after rename
infoStreams.get(0).setStepname("Old Source");

// after: reference the actual connected step
infoStreams.get(0).setStepname("Sorted Customer Rows"); // step that actually hops into Merge Join

// before execution, verify both inputs are connected
for (StreamInterface stream : mergeJoinMeta.getStepIOMeta().getInfoStreams()) {
  StepMeta source = transMeta.findStep(stream.getStepname());
  if (source == null || !transMeta.isStepUsedInHops(source)) {
    throw new IllegalStateException("MergeJoin input step missing/disconnected: " + stream.getStepname());
  }
}

boolean hasConnectedInput(TransMeta transMeta, StreamInterface stream) {
  StepMeta s = transMeta.findStep(stream.getStepname());
  return s != null && !s.isDisabled() && transMeta.isStepUsedInHops(s);
}

try {
  processRow();
} catch (KettleException e) {
  if (e.getMessage().contains("UnableToFindSpecifiedStep")) {
    logError("Reconnect the missing MergeJoin input hop in Spoon and re-run", e);
  }
  throw e;
}

When it happens

Trigger: The step listed first in the MergeJoin meta's info streams has no connected hop into the Merge Join step — hop deleted or renamed, step name mismatch, or the transformation was executed with a disabled/removed source step.

Common situations: Renaming the first input step without re-opening the Merge Join dialog (hop references stale names); copying a Merge Join step into another transformation without reconnecting inputs; disabling a source step during debugging; partial transformations assembled programmatically.

Understand the failure class

Background: 'Could not be found', 'does not exist', 'not found in database': the resource-not-found family when an ID, slug, key, or URI lookup comes back empty — this error's family across 20 libraries.

Related errors

• A deadlock was detected between steps
• BaseStep.Exception.SourceStepToReadFromCantRunInMultipleCopies
• BaseStep.Exception.TargetStepToWriteToCantRunInMultipleCopies
• BaseStep.Exception.SourceStepToReadFromDoesntExist
• BaseStep.Exception.TargetStepToWriteToDoesntExist

AI-assisted analysis of pentaho/pentaho-kettle@f3058517a1 (2026-09-13). Data as JSON: /api/errors/b1aea08c6c3f31ae. Report an issue: GitHub.

Appendix: source

Thrown at engine/src/main/java/org/pentaho/di/trans/steps/mergejoin/MergeJoin.java:76

    Trans trans ) {
    super( stepMeta, stepDataInterface, copyNr, transMeta, trans );
  }

  public boolean processRow( StepMetaInterface smi, StepDataInterface sdi ) throws KettleException {
    meta = (MergeJoinMeta) smi;
    data = (MergeJoinData) sdi;
    int compare;

    if ( first ) {
      first = false;

      // Find the RowSet to read from
      //
      List<StreamInterface> infoStreams = meta.getStepIOMeta().getInfoStreams();

      data.oneRowSet = findInputRowSet( infoStreams.get( 0 ).getStepname() );
      if ( data.oneRowSet == null ) {
        throw new KettleException( BaseMessages.getString(
          PKG, "MergeJoin.Exception.UnableToFindSpecifiedStep", infoStreams.get( 0 ).getStepname() ) );
      }

      data.twoRowSet = findInputRowSet( infoStreams.get( 1 ).getStepname() );
      if ( data.twoRowSet == null ) {
        throw new KettleException( BaseMessages.getString(
          PKG, "MergeJoin.Exception.UnableToFindSpecifiedStep", infoStreams.get( 1 ).getStepname() ) );
      }

      data.one = getRowFrom( data.oneRowSet );
      if ( data.one != null ) {
        data.oneMeta = data.oneRowSet.getRowMeta();
      } else {
        data.one = null;
        data.oneMeta = getTransMeta().getStepFields( infoStreams.get( 0 ).getStepname() );
      }

      data.two = getRowFrom( data.twoRowSet );

View on GitHub (pinned to f3058517a1)