Facing problem in implementing the analytical query in OWB

Question

Hi All,
I was trying analytical function for my Customer Dimension but am not able to implement the same in my OWB.I have 2 tables who have duplicates customers which cannot be figured out by Customer id as customer id is not unique.So, i have to filter out the information on the basis of First Name, middle name, last name & DOB. Since, each customer has multiple address records( for Permanent, Correspondence, Unknown ).
I want 1 record for all the three address types.
Can i do thru the following query???
If yes, how will i implment the same in OWB.
SELECT PRENOM, MIDDLE_NAME, NOM, DEBUT, SEXE, OCCUPATION, FAMILY_STATUS,
     EDUC_QUALIFICATION, CUSTOMER_NATIONALITY, SIGLE, ANNUAL_INCOME,
     FATHER_HUSBAND_NAME,TYPE_ADDRESS, ADDRESS1, ADDRESS2, ADDRESS3,
     TOWN, REGION, POSTCODE,ADDRESS_EFFECT_DATE
FROM
(SELECT C1.PRENOM, C1.MIDDLE_NAME, C1.NOM, C1.DEBUT, C1.SEXE, C1.OCCUPATION, C1.FAMILY_STATUS,
     C1.EDUC_QUALIFICATION, C1.CUSTOMER_NATIONALITY, C1.SIGLE, C1.ANNUAL_INCOME,
     C1.FATHER_HUSBAND_NAME,C1.TYPE_ADDRESS, C1.ADDRESS1, C1.ADDRESS2, C1.ADDRESS3,
     C1.TOWN, C1.REGION, C1.POSTCODE,C1.ADDRESS_EFFECT_DATE,
ROW_NUMBER() OVER (PARTITION BY C1.PRENOM, C1.MIDDLE_NAME, C1.NOM, C1.DEBUT,C1.TYPE_ADDRESS
ORDER BY C1.ADDRESS_EFFECT_DATE DESC)M
FROM CUST_1 C1
UNION ALL
SELECT C1.PR_FIRST_NM, C1.PR_MIDDLE_NM, C1.PR_LAST_NM, C1.PR_DOB, C1.PR_GENDER, C1.PR_OCCUPATION, C1.PR_MARITAL_STATUS,
C1.PR_EDUCATION_QUAL, C1.PR_NATIONALITY, C1.PR_TITLE, C1.GP_ANNUAL_INC, C1.PR_FATHER_NM,
C1.AD_TYPE, C1.AD_ADDR1, C1.AD_ADDR2, C1.AD_ADDR3, C1.TOWN, C1.REGION, C1.AD_PIN_CD,C1.REC_UPDT_DT,
ROW_NUMBER() OVER (PARTITION BY C1.PR_FIRST_NM, C1.PR_MIDDLE_NM, C1.PR_LAST_NM, C1.PR_DOB,C1.AD_TYPE
ORDER BY C1.REC_UPDT_DT DESC )M
FROM CUST_2 C1)
WHERE M = 1
Please help me out as right now, am using aggregators which are taking lot of time. Loading type on my Customer Dimension table is- Insert/Update
Thanks in advance
-Nikita.

Answer 1

Hello Nikita,
From your first post I understand that you would like to merge 3 similar (but not complete identical) records into 1 record. If this is the case, then have you thought about using the match-merge operator in OWB?
With the match-merge operator you can tell the mapping which records are similar by selecting "first_nm", "middle_nm", "last_nm" in the Match-Bin of the operator.
Then you define how attributes should be merged.
This way you do not need to aggregate your source data set. Just select all records from both source tables. Put those in the Set-Operator "UNION ALL" and this will be you input for the match-merge operator.
Regards,
Ilona