In general, it depends on the arrival order of rows. In the case of using it with Beam, the order of rows is lost. But you can work around it with Memory Group By step. See: https://hop.apache.org/manual/latest/pipeline/beam/getting-started-with-beam.html#_unsupported_transforms
Also... Would it be helpful if the docs were edited to say perhaps for Row Denormaliser something like "If using with `Group fields` option, ensure that your input is sorted" ??? Thad https://www.linkedin.com/in/thadguidry/ https://calendly.com/thadguidry/ >