如何在 Postgres 中为文本定义正则表达式

Question

Please help to define Postgres regexp for this case:

I have string field:

union all select 'AbC-345776-2345' /*comment*/ union all select 'Fgr-sdf344-111a' /*BN34*/ some text union all select 'sss-sdf34-123' /*some text*/ some text

Here is the same text in select statement for convinience:

select 'union all select ''AbC-345776-2345'' /*comment*/ union all select ''Fgr-sdf344-111a'' /*BN34*/ some text union all select ''sss-sdf34-123'' /*some text*/ some text' as str

I need to get from this mess text only values in '...' and select it into separated rows like this:

AbC-345776-2345
Fgr-sdf344-111a
sss-sdf34-123

Pattern: 'first 2-3 letters - several letters and numbers - several letters and numbers'

I created this select but it contains all comments and "sometext" as well:

select regexp_split_to_table(trim(replace(replace(replace(replace(t1.str,'union all select',''),'from DUAL',''),chr(10),''),'''','') ), E'\\s+')
from (select 'union all select ''AbC-345776-2345'' /*comment*/ union all select ''Fgr-sdf344-111a'' /*BN34*/ some text union all select ''sss-sdf34-123'' /*some text*/ some text' as str) t1; 

Answer 1

The following should do it:

select (regexp_matches(str, $$'([a-zA-Z]{2,3}-[a-zA-Z0-9]+-[a-zA-Z0-9]+)'$$, 'g'))[1]
from the_table;

Given your sample data it returns:

regexp_matches 
---------------
AbC-345776-2345
Fgr-sdf344-111a
sss-sdf34-123  

The regex checks for the pattern you specified inside single quotes. By using a group (...) I excluded the single quotes from the result.

regexp_matches() returns one row for each match, containing an array of matches. But as the regex only contains a single group, the first element of the array is what we are interested in.

I used dollar quoting to avoid escaping the single quotes in the regex

Online example