duplicate elimination of master records (customer, vendor)

[former_member193516]

Hi all,

we have a problem in master record management that there are about 20'000 records in each customers and vendors. Some of them are certainly duplicates. They could be fairly identical in names and address or similar.

Now it has been requested that those duplicates need to be identified and for future use in regular time slots checks should be done to deal with those (clear open items, block duplicates etc.). Excel as a first help is no option...

I could imagine having a few objects only for checks, for example:

If

1. Country same

2. City same

3. check Name - maybe 5 letters then write both identified entries into table with additional fields from master record (name2, post code, telefon etc) and give out ALV protocol so user could go into further analysis

Any further ideas?

Thanks

Hein

[Former Member]

Hi,

If you are ABAPer follow the below procedure or Take ABAPer help. he will help you.

Procedure:

retrive data from data base table into internal table based on update date (Field name: aedtm ) and time in descending order.

pass those records into ALV FM and display data as in required format.

(or) use delete adjaucent duplicates keyword.

it will helps you definitely.

Ram.

[former_member193516]

Thanks.

Tables we would need to look at are kna1 for customers and lfa1 for vendors.

Regards

Hein

[Former Member]

I'm working on a similar requirement. What we have done is set up a ranking of matches by how well they match. For example, if the name is an exact match, the priority is 100; if the address and first letter of the postal code matches, the priority is 80 and so on.

It's quite complex because the address is essentially entered freeform.

Good luck.

Rob

[brad_bohn]

How are your address sort keys maintained or are they maintained? SORT1 and SORT2 are typical starting points for duplicate checking with address data if you maintain one of the fields with a combination of (partial) name, city, zip or some other identifiers.

I use the same approach as Rob though...a probability of a duplication based on the degree of a match on the address fields, email address and phone numbers. It gets difficult when you have different city name spellings that are acceptable for the same city or name abbreviations vs. non-abbreviated names.

[Former Member]

you can use select distinct in sql statement than you can avoid duplicate records

thanking you

[brad_bohn]

you can use select distinct in sql statement than you can avoid duplicate records

I often wonder if you're reading posts at all - your responses often have nothing to do with the issue in question :-S

[Clemenss]

Hi Hein,

some useful hints have been given. Rob's approach is useful, I think this is a common task frequently ignored by SAP standard. I remember we used Function module RS_COMPARE_WORDS_SIMILAR to identify duplicates in street names - this is just the function used by ABAP editor error to give hints for similar expressions.

Anyway: You should implement a user exit for the creation of new business partners / vendors / customers that searches for similar or equal database entries and give a warning.

[Duplicate Check|http://help.sap.com/saphelp_nw73/helpdata/en/09/24a63a27225b4ae10000000a11402f/frameset.htm] is an integrated part of Business [Address Services (BC-SRV-ADR)|http://help.sap.com/saphelp_nw73/helpdata/en/c8/13b237b9a9a968e10000009b38f8cf/frameset.htm].

I think it is worth to check out this and the respective BAdI definitions and function module ADDR_FUZZY_SEARCH.

Regards,

Clemens

[former_member193516]

programing will be used.