REGEXP_REPLACE - two letters in a string

Question

I have a string:

string := '</br>, <br/>, \n<br>, jane, brutus'

I have to replace all 'br' marks by '/n' mark. I need an output like this:

'/n, /n, n/, jane, brutus'

I tried:

select regexp_replace('</br>, <br/>, \n<br>, ania', '^.*((br)[^,]+)', '\n', 1) from dual;

But then output is like this:

'/n'

Any ideas how can I resolve this problem?

I mean I tried: select regexp_replace(',
, \n
, ania, brutus', '^.*((br)[^,]+)', '\n', 1) from dual; — Mateusz Srebniak, Aug 01 '22 at 06:39

Thorsten Kettner · Answer 1 · 2022-08-01T07:31:51.747

0

It is easier than you thought. Just look for the 'br> parts. This is: an opening angle bracket, an optional slash, the letters b and r, an optional slash, a closing angle bracket: </?br/?>.

The whole expression:

select regexp_replace('</br>, <br/>, \n<br>, ania', '</?br/?>', '\n') from dual;

If you want to replace all groups of adjacent marks and treat br marks the same as \n in your search, then it gets a little bit more complicated:

regexp_replace('</br>, <br/>, \n<br>, ania', '(\\n|</?br/?>)+', '\n')

edited Aug 01 '22 at 07:31

answered Aug 01 '22 at 06:41

Thorsten Kettner

89,309
7
49
73

It doesn't work with '\n
'. Output is: \n, \n, \n\n, ania – Mateusz Srebniak Aug 01 '22 at 06:46
I replace all br marks with \n. What exactly do you want differently? – Thorsten Kettner Aug 01 '22 at 06:48
I mean '\n
' is also one mark with 'br'. I need output like '\n' but with your query output is '\n\n' for this one mark – Mateusz Srebniak Aug 01 '22 at 06:54
Can we generalize: replace any group of br/n marks wth one \n? I.e. '
\n
hello\n\n

' would become '\nhello\n'? – Thorsten Kettner Aug 01 '22 at 07:00
I need '\n' mark instead of every comma separated 'br' mark – Mateusz Srebniak Aug 01 '22 at 07:04
I am not sure I understand. Maybe we mean the same thing. I have updated my query with the generalized "replace any group of br/n marks with one \n". – Thorsten Kettner Aug 01 '22 at 07:11
1

@JvdV: Thank you. You are right of course. I've updated my answer. – Thorsten Kettner Aug 01 '22 at 07:32

Motif · Answer 2 · 2022-08-01T12:08:15.890

0

If I understood correctly this code will be work.

SQL>
SELECT word,
   REGEXP_REPLACE (REGEXP_REPLACE (word, '([\\<])([a-zA-Z/]+) 
                    ([>])', '/n'),
                   '\\n([a-zA-Z/]+)',
                   '\1')
      AS Result
   FROM (SELECT '</br>, <br/>, \n<br>, jane, brutus' AS word FROM DUAL)

RESULT

/n, /n, /n, jane, brutus

SQL>

edited Aug 01 '22 at 12:08

answered Aug 01 '22 at 11:57

Motif

69
9

REGEXP_REPLACE - two letters in a string

2 Answers2

RESULT