Is there a book that...

[jjoensuu]

Hi all,

I am looking for a book, web page or white paper that discusses issues of a encountered when updating one database from another. Or when trying to reconcile data from a source (or multiple source) with data in a target system.

The issues that I am interested in are of the unexpected abstract type, pertaining to factors such as:
-data in a field having a different meaning (but same value) in certain cases
-lack of information on why data may have changed (or is missing)

Apologies in advance if my question is too vaguely formulated. I'm in a lack of good examples at the moment.

-JJ

 

[ceh4702]

Finding multiple SSN's is always fun. If you ever used the SSN as a key, it can cause problems. There are always fictitious ssns that have been deleted and then ssn's that are missing.

Our school system imports Electronic FAFSA (Fed appl for student aid), and the system uses a waited results system looking for a match on the last name the SSAN and the Date of birth. You get so many points for a match and then a near miss in the different categories and then you define what you will accept as a match.

This new system has multiple addresses and multiple phone numbers at each address.

One big thing we had was an outside firm that helped us to get this off of the ground. It is not fun going to a completely foreign and new system. Requires lots of meetings to decide what kind of naming standards you will accept. Then there are val code tables, and all sorts of naming conventions you may want to consider. Sometimes you have to force the managers to make a decision, and move the project along.

We had old programmers working on extraction porgrams and new programmer/consultants working on import programs. Sometimes you have to make compromises about how much data to import and what not to import.

Then there are reporting issues. It is hard to do and end of year report if half the files are on one system and half are on another system.

We comparmentalized our projcect a bit into major system categories like Accounting, purchase orders, Payroll, HR, Student System, Registration, Etc. We changed so many ways we did business that it was hard to keep up. Plus the new system did some things completelly different so the business practices had to be redesigned.

People are stubborn and then try to get the same reports they use to use on the new system even though the new system is run completely different. The end-reports coming out are more important than the data. Making all new reports for a new system can be a lot more difficulty than moving the data.



If you do not like my post feel free to point out your opinion or my errors.