Hi Sania G.Nitish J.Sonal G. does Zingg supports below scenario, matching the records between 2 sources based on say 3 columns. It should first try to match where all 3 columns have values. Then from remaining records if 2 columns are present of source1 side it should take those for matching and lastly for records having values in one column only. This is for comparing 2 sources and identifying the match records based on 3 columns but there are scenarios where data could be present in all 3 columns and sometimes absent from 1 or more columns. But priority is wherever values are present it should match them first and then the null records.
But can we define the rule for that...try with the all 3 columns first, then match for 1st and 2nd column values, at the end 1st and 3rd column...or all this is part of the label phase and the model is to be trained that way?