on 11-17-2015 1:19 PM
How would I be able to achieve the following (I tried using a classic chain of a table compare, history preserve and row generation, but to no avail, as only the most recent record gets updated):
Suppose I have an employee dimension with just five columns: surrogate key, business key, employee name, position, date of birth. I only want inserts of new records and changed records based on a change of function, i.e. the other columns I'm not really interested in in terms of history, so they should really just be overwritten.
So if an employee has changed position a number of times, but HR finds out that the year of her DOB was initially entered incorrectly I want all of the records pertaining to that person to be corrected. How would I be able to achieve this (ideally the whole logic in one single data flow)? Thanks in advance.
You can splitthe flow after the History_Preserving transform in an Update and Insert path. Use two Map_Operation transforms to that extent.
Join the Update path with the original table, overwrite those columns you don't want to keep history for and use another Map_Operation to switch the rowtype to Update again.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
User | Count |
---|---|
101 | |
13 | |
13 | |
11 | |
11 | |
7 | |
6 | |
5 | |
4 | |
4 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.