[Metafacture] SUM algorithm & grouping questions

Nico Siebert nsiebert at avantgarde-labs.de
Tue Jun 17 13:03:33 CEST 2014

Hello everyone!

I have some questions about the metafacture framework.

Are there any more information regarding the SUM algorithm apart from 
the flowchart?
I understand that for each record 3 matchkeys are created and those will 
be compared to the corresponding matchkeys of another record.

Now I wonder:

1. Is *every* key of a record compared to the corresponding key of a 
second record? If so, will any match result in a group or must at least 
2 MatchKeys be equal? What is the logic behind this?

AMatchKey1        compareTo    BMatchKey1    RESULT=false
AMatchKey2        compareTo    BMatchKey2    RESULT=false
AMatchKey3        compareTo    BMatchKey3    RESULT=true

--> Results in grouping of Record A and B ?

2. Is there any weighting implemented or are the matchkeys simply 
compared like mk1.match(mk2)?

Thank you and kind regards
Nico Siebert

Nico Siebert
Intern | Project SLUB/DMP

Telefon: +49(0)351 - 215 908 34
Fax:     +49(0)322 - 241 80384

Email: nsiebert at avantgarde-labs.de
Website: www.avantgarde-labs.de
Avantgarde Labs GmbH | Löbauer Straße 19 | D-01099 Dresden

Geschäftsführung: Robert Glaß, Torsten Hartmann, Sandy Lucka, Sven Rega
Sitz Dresden | Amtsgericht Dresden | HRB 31215

More information about the Metafacture mailing list