2 Replies Latest reply: Oct 29, 2013 8:27 AM by Shreyas Ram R RSS

    Record filter Duplicates Endeca

    Shreyas Ram R

      Hi Friends,

       

      Please help us in avoiding duplicates in the following scenario.

       

      I have a product carried out in 2 different wine groups example group "A" and "B" but if the product is available in both groups,then I need to consider the record from group B.

       

      I am using the following record ilter which will definitely yield 2 records.

       

      AND(Record_Type:Wine,OR(group:A,group:B))

       

      Please advise how to proceed with this requirement.Due to some constraints,we cannot change the architecture and we need 2 records to be maintained i.e per group we need  record.

       

       

      Regards

      Shreyas

        • 1. Re: Record filter Duplicates Endeca
          Vaibhav R Patil

          Hi Shreyas,

           

          It’s not very clear what exactly you mean by groups here. It can be looked from different perspective, following are the various possibilities and the potential solutions

           

          Number 1 :- Lets say a product is part of group A and group B and you are exposing group dimension. In this case, do you want to display product when group --> B is clicked and not when group --> A is clicked?

           

          This scenario will need to be handled while you are ingesting data to Endeca. Simplest option will be to add record cache into the pipeline and configure it avoid duplicate association to A & B if you get two records for this association. If there is any complex logic than a simple de-duping then  you can create a data manipulator (java manipulator) for the same.

           

          Number 2 :- Are you using experience manager? and do you want have a functionality of displaying records with B before than records with A, then you case use boost and bury feature to display records associated with b first.

           

          Number 3:- If you have two records then you can do a record roll up to get both the records a single entity. If both products have a same product id then you can roll up on the product id. Then you can display aggregated refinements counts as well in the facets to display count. This will fix your both problems. But you will have to be bit careful with this a record rollup increase the response size, if you choose to return all records in the rolled up entity instead of a representative record

          Hope that helps

           

           

          Many Thanks,

          Vaibhav R Patil

          • 2. Re: Record filter Duplicates Endeca
            Shreyas Ram R

            Thanks Vaibhav for the response.

             

            PFB my responses.

             

            Number 1-Actually group is not a dimension and it is just a property.

            PFB an example:

             

            Record-1

            product_id:prod12

            group:A

            Online_Inventory:instock

            Record_type:Wine

             

            Record-2

            product_id:prod12

            group:B

            store_Inventory:instock

            Record_type:Wine

             

            My record filter which I've specified here fetches me both the products because of which I'm seeing duplicates and why we are following this approach is there are few properties which are present only in group_A record which will not be present in group_B,hence when a record is carried out in both groups,we need the record from B group.

             

            Number 2:No,we donot have Experience Manager.

             

            Number 3:-If I perform a roll up,I can get only 1 record but can you please tell me if there are unique properties in 2 records,can I get all those in the same record?Then,I feel,we can roll up the records which will fix this issue.

             

            Regards

            Shreyas