We Don't Have Clean Data

I have an SSIS package.

It's a nice package.

Has two containers, each container produces a single text file.

So I finished coding the changes a few months ago.

And packaged the files up for the client to test.

And their people were on vacation.

So we waited.

Files failed.  Said the file lengths were off.

Hmm.  They don't seem off to me.  QA lady agreed.

So we sent them back.


Records already exist in their test system.

Hmmm.  Let me modify the Reference number to higher range which would not be in their system.

Package files, sent over, their people are again on vacation.

Failed.  Again.

The inserts went in fine but the updates failed, records don't exist in their system.

Hmmm.  Modified the Updates to retain their existing number.

Sent files, again.

Failed.  Data does not match what's expected.

Hmmm.  I only modified field lengths, none of the business logic changed.

So I responded back, as best I could, without showing signs of frustration, that we have specific requirements which we must package the data, which we learned from trial and error, and the data was are sending is not clean, in the fact that all the data we're sending already exists on the client database, so we have to fake it out.  In addition, because we don't have clean data, I'm manually deleteing records for the sole purpose of creating data, which otherwise would not exist, because we have no data that conforms to the required standards.

And the business analyst says, "Can the contractor produce a file that works?".

Hmmm.  This is getting to be too much.