Regex to replace substring occurrences only when they are enclosed between begin / end brackets

Question

regex problem in ruby, and for some reason, I need this in one line in gsub method of ruby

Suppose the input sample string variable is multi line string like below

begin1 item abc item abc item 
extra end1
begin2 item abc item abc extra end2
begin1 item abc item abc extra end1

The rule is to change all item which inside block begin1 and end1 into love, it may across multi lines

After replace, the output sample should be

begin1 love abc love abc love 
extra end1
begin2 item abc item abc extra end2
begin1 love abc love abc love end1

The solution is something like this

puts sample.gsub!(/(begin1.*)item*(.*end1)/m,'\1love\2')

Alexis · Accepted Answer · 2013-06-15T23:14:06.877

3

What about this one: sample.gsub!(/item(?=((?!begin1).)*end1)/m, 'love')?

And, to make the regex a little bit less magic:

/ 
  item              
  (?=               # look-ahead assertion
    (               
      (?!begin1)    # negative look-ahead assertion
      .             # any character that is not part of the 'begin1' sequence
    )*              # there's no 'begin1' after 'item' and before 'end1'
    end1            # there is 'end1' after 'item'
  )
/mx                 # /m - for multiline strings, x - for comments

(Note, that it will work correctly only if your 'markup' is 'correct' - e.g. there is no trailing ends or leading begins, no nesting blocks)

Also, I can't help to say that regexes is poor tool for such a task. Complex regular expressions are too incomprehensible and too unmaintainable, they are like black magic - you never know when they're gonna crash :)

edited Jun 15 '13 at 23:14

answered Feb 02 '12 at 21:30

Alexis

4,317
1
25
34

So magic, and it works exactly what I want. And it works fine for small input, when it goes big, it reports `Stack overflow in regexp matcher` – Larry Cai Feb 03 '12 at 01:47
the answer is good enough, `stack overflow` is related with ruby 1.8.7 only, solution works fine in 1.9. thank alexis. see real script https://github.com/larrycai/sdcamp/commit/923237e5c – Larry Cai Feb 03 '12 at 05:08
I tested it with 100M long string - it took pair minutes to finish, but there was no 'stack overflow' in 1.8.7 – Alexis Feb 03 '12 at 10:15

score 0 · Answer 2 · answered Feb 02 '12 at 16:02

I'm not entirely certain what you mean with the input part of your question, so I'm a bit confused but I think you're using the gsub method the wrong way. Try this:

> string = 'begin1 item abc item abc item'
=> "begin1 item abc item abc item"
> string.gsub(/item/, 'love')
=> "begin1 love abc love abc love"

It's a simple regular expression substitution method. Read the docs for more details on how to use it. I think you may be making it more complex than it needs to be.

I update the question. the input is the multi line string, and change `item` between `begin1` and `end1` to `love` in one line — Larry Cai, Feb 03 '12 at 00:36

Regex to replace substring occurrences *only* when they are enclosed between begin / end brackets

2 Answers2

Regex to replace substring occurrences only when they are enclosed between begin / end brackets