Anti-Join in Paxata

Anti-Join in Paxata

Hi again everyone!  You've got to be sick of hearing from me by now.

So on the continuing monster project I've been working on, the SPSS stream that I am to match/mimic in Paxata calls for a a merge between two data sets on four keys columns.  But it's an "anti-join" between two data sets, to "include records in the first data set that do not match any others (anti-join)."  Looking at the options for Look-up, I don't see a way to do this.  Would this be done via Append instead or is there something I'm missing?  Any thoughts appreciated.  
Labels (1)
2 Replies

From what you are describing this can easily be accomplished in two steps.  First do the join between the two datasets.  The default options (on the options tab in the lookup definition) will show a left outer lookup.  Don't change this.  After the join is executed, bring up the filtergram on the "Sources" column.  It is the very first column of data.  This system managed column will show you which records came from each dataset.  Filter for the row that shows data from both datasets, then click on the remove tool and save it.  You will be left with only the records that were in the initial dataset that are not in the second dataset.

Okay, that makes sense and seems very approachable.  Thank you!