For a while I’m being involved in discussions about ETL code review issue and I’m glad to share my understanding of it with everyone here.
First of all, I see that ETL code review is a live issue and people understand the importance of standardised ETL code though come to the issue from different perspective.
As for me, I seeETL code has much in common with a wardrobe.
Imagine your wardrobe. If you keep all your staff in order, how long will it take you to find your favourite pair of red socks? I reckon, a few seconds or so.
And now imagine you have a mess of your clothing in there. How long will you be looking for your socks in this case? And what about your girlfriend if you ask her for such a favour? And what if your wardrobe is 10 times bigger? Imagined? The same is with ETL code produced.
If there are no standards of ETL code quality and no processes to trace it… if there is, but you don’t enforce it or simply skip it under pressure of project deadlines… soon it will become a problem to find anything there.
It will impact your system further support and modification. When a problem occurs, it takes time to find the cause of it…quite a lot of time. Still worse when someone developed the code for you. In this case, you are that girlfriend looking for a pair of red socks in a huge messy wardrobe.