Matching 2 conditions in regexp_replace in postgres

Question

I need help getting my regex syntax right in Postgres regexp_replace: My string:

1ABC 2ABC 3DEF 4DEF

My 2 Match/Replace conditions are:

 Replace: A OR C but not in front of 2
    or
 Replace: D OR F but not in front of 4

So I'm expecting to get:

"1A;BC; 2ABC; 3D;EF; 4DEF;"

My partial replace for condition 1 is:

SELECT regexp_replace('1ABC 2ABC 3DEF 4DEF', '((?<!2)(A|C))','\1;','g' );

My 'Replace' is really an 'Insert' after the matched literal.

I just can not seem to be able to find the pattern for the 2nd condition, without breaking the whole thing. Is this even possible in 1 statement?

404 · Accepted Answer · 2018-04-24 12:41:45Z

0

SELECT regexp_replace('1ABC 2ABC 3DEF 4DEF', '((\w*[^2]A|C)|(\w*[^4]D|F))','\1;','g');

Result: 1A;BC; 2ABC; 3D;EF; 4DEF;

answered Apr 24, 2018 at 12:41

404

8,7622 gold badges34 silver badges53 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

And eurotrash is a genius! :-) Thanks!