Diff on Sorted Keys

Diffs two sorted streams of rows

Processing

The step accepts two streams of rows, coming from two input gates in and reference.

Input rows are matched to reference rows based on key fields. The step uses a diffing algorithm that requires both row streams to be be sorted on key fields in ascending order with nil values last, if any.

• if an input row has a matching reference row, it is either identical or changed depending on whether the data fields match as well
• if an input row has no matching row in the reference stream, it is considered new
• if a reference row has no matching row in the input stream, it is considered deleted

If the key fields in either stream are not sorted or contain duplicates, the behaviour of the step is undefined.

Settings

Results

{
  :<ref data field name> {
    :input <value on input side>,
    :reference <value on reference side>
  },
  ...
}