Page 1 of 1
Posted: Wed Feb 13, 2008 4:10 pm
by vmcburney
An SCD type 2 table is supposed to hold duplicates - it shows changes to that particular CUST_ID over time. When the REC_CURRENT_IND flag is set to Y that is the current record - that's the one used by most reports and queries. When the flag is set to N that's a superceded record and the REC_EFFDT field shows you when that row came into effect. I have only seen SCD Type 2 tables with START and END date fields so you can use between joins for historical reporting. So in this case there should also be a REC_ENDDT field set to a high date for current records (2999-12-31) and set to an end date for superseded records.
Posted: Wed Feb 13, 2008 6:55 pm
by kduke
Link does not work.
Posted: Wed Feb 13, 2008 7:43 pm
by chulett
Remove the extra dot from the end. After you are done poking around over there, don't forget to wash your hands.
Posted: Wed Feb 13, 2008 10:03 pm
by ray.wurlod
Start at
this page for the full set. Heed Craig's advice.
Posted: Thu Feb 14, 2008 5:03 am
by ray.wurlod
Of course it can. It's the surrogate key that can not. If some attribute of the customer (maybe name) changes, in a Type 2 SCD there will be a new row with the same CUST_ID, but with a new surrogate key.
Posted: Tue Aug 05, 2008 6:30 am
by laknar
Hi,
I have exactly the same design.
When we load the records first time into the dimension table.
How can i load without Suurogate key stage,trigger.
Posted: Tue Aug 05, 2008 6:43 am
by chulett
Start a
new post. Include the details of
your job and state
your problem there.