Re: Using SQL, replace all characters less than x'40' in a string with blank

On 10/23/2018 11:45 AM, Dan wrote:

Thanks Buck! That is where I was headed if I couldn't do this in native

SQL.

I'm not going to preach; everyone has their own situation they have to
deal with. I should have been forthcoming with my reasoning though. I
have done very similar things; written a routine in RPG, deployed it
into who knows how many programs [1], and then wanted similar
functionality in SQL, so I wrote it in SQL, and deployed that into who
knows how many places.

And then there was a legitimate need to change the way that routine
works. Imagine, in your example, that you also want to translate high
EBCDIC values to upper as part of this text cleanup.

With two separate code blocks, deployed in many hard-to-find places,
that simple change is now a lot of work. If I'd encapsulated the single
RPG sub-procedure into an SQL UDF then all consumers of that business
function would have been handled simply.

That's a lot of words to explain how I always seem to regret multiple
implementations of the same business function.

[1] The lack of built-in cross reference tooling is probably my only
genuine gripe with the platform.