Z_cluster id changes after running Link job

[iitbhumanish]

Hi ,

Just need quick information after running Link job. I see that Z_Cluster id is changed after link job , it did not use same Cluster id created while running Match job for the same match records.

Is there a way to get LINK job to use same cluster id generated through MATCH job ? Please advise.

For ex: four duplicate records Stewart were kept under Cluster Id = 8305 after MATCH job

As part of Incremental load , it generates different cluster id for another incoming duplicate Stewart after LINK job .

[sonalgoyal]

can you please tell me what are the input datasets you are passing to the link phase?

[iitbhumanish]

I am passing both Exist and Delta as input. Please refer below screen shots

Note : For existing record, I had converted matched records ( generated through MATCH job) into temp file without three extra fields : minScore, maxScore and ClusterId

[sonalgoyal]

Got it, thanks. In Zingg OSS, link is a different phase from match and designed to find record similarity across datasets and not designed for incremental processing post a one time match. Hence you see the results you are getting.

[iitbhumanish]

Thanks Sonal for the response.