Regex: I want to modify something

[guy038]

Hello Vasile,

The most simple regex, to select the lines, which contains the word WORD_1 OR the word WORD_2, seems to be :

• .*(WORD_1|WORD_2).* for the line contents, only

• .*(WORD_1|WORD_2).*\R for the complete lines, with their line-breaks

---

Seemingly, as you want to delete these lines, the correct S/R, that you need, is :

SEARCH .*(WORD_1|WORD_2).*\R

REPLACE EMPTY

Best Regards,

guy038

[Vasile Caraus]

hello guy038. No no.

I want to delete everything before the line that has WORD_1 included that line. And to delete everything after WORD_2 included that line. So as to remain only those 3 lines with “my_text.”

The problem of my regex is that selects everything before WORD_1 and everything after WORD_2 but it doesn’t select also the rest of the lines which contains this 2 words.

Please see again my regex and my example.

[guy038]

Vasile,

Ah, OK ! Now, I understand what you want to get :ONLY text, even in some lines, between the line, containing the string WORD_1, excluded and the line, containing the string WORD_2, excluded, don’t you ?

Well, a solution could be the S/R, below :

SEARCH (?-s).*WORD_2(?s).*|.*WORD_1.*?\R

REPLACE EMPTY

This regex would delete any possible standard character before the string WORD_2 and every text after that string OR any possible character ( standard or EOL ) before the string WORD_1 as well as the smallest range of character(s) till an End of Line character( i.e. the remaining characters of the line, after WORD_1, including its line-break )

So, from the original text :

Line 01
Line 02
Line 03
Line 04
Line 05
----- WORD_1 -----
Line 07
Line 08
Line 09
----- WORD_2 -----
Line 11
Line 12
Line 13
Line 14

We get the final text, below :

Line 07
Line 08
Line 09

NOTES :

• By inverting the two terms of the alternative |, it prevents me to add the an extra (?s) syntax, at beginning of the second part of the alternative !

• Beware, this regex is correct, ONLY IF there’s a SINGLE couple WORD_1 - WORD_2, in your file !!

---

In case of multiple couples of lines, containing WORD_1 and WORD_2, in your file, an other S/R is necessary :

SEARCH (?-s).*WORD_2(?s).*?(WORD_1.*?\R|\z)|((?!WORD_2).)*?WORD_1.*?\R

REPLACE EMPTY

Then, given the original text, below :

Line 01
Line 02
Line 03
----- WORD_1 -----
Line 05
Line 06
Line 07
Line 08
Line 09
----- WORD_2 -----
Line 11
Line 12
Line 13
Line 14
Line 15
----- WORD_1 -----
Line 17
Line 18
----- WORD_2 -----
Line 20
Line 21
Line 22
Line 23
Line 24
----- WORD_1 -----
Line 26
Line 27
Line 28
Line 29
----- WORD_2 -----
Line 31
Line 32
Line 33
Line 34

This second S/R would produce, as expected, the changed text, below :

Line 05
Line 06
Line 07
Line 08
Line 09
Line 17
Line 18
Line 26
Line 27
Line 28
Line 29

NOTES :

• The first part (?-s).*WORD_2(?s).*?(WORD_1.*?\R|\z) looks for any block of text between, either :

  • The beginning of a line, containing the string WORD_2 and the end of the line, containing the nearest string WORD_1

  • The beginning of a line, containing the string WORD_2 and the very end of the file

• The second part ((?!WORD_2).)*?WORD_1.*?\R, with an implicit modifier (?s), looks for any block of text, which does not contain the string WORD_2, located between the very beginning of the file and the end of the line containing the nearest string WORD_1

---

Remark : To be a bit more restrictive about the key words WORD_1 and WORD_2, you may add the modifier (?-i) to forces a sensitive searching, which, therefore, changes the two regexes, above, into :

SEARCH (?-is).*WORD_2(?s).*|.*WORD_1.*?\R

SEARCH (?-is).*WORD_2(?s).*?(WORD_1.*?\R|\z)|((?!WORD_2).)*?WORD_1.*?\R

Cheers,

guy038

P.S. :

Of course, these regexes suppose, also, that NO line, of your file, may contain, simultaneously, the two key-words WORD_1 and WORD_2 !!

---

To end with, here is an other form of the second S/R, with non-capturing groups. Could be faster, in case of a huge file !

SEARCH (?-is).*WORD_2(?s).*?(?:WORD_1.*?\R|\z)|(?:(?!WORD_2).)*?WORD_1.*?\R

REPLACE EMPTY

[Vasile Caraus]

Thanks guy038, all your Regex works fine. But sometimes I have a problem with \R sequence. Sometimes works, sometimes doesn’t work. I write about this problem in other topics.

So, I find another 2 solution, without using \R sequence.

(?s)((^.*)(WORD_1).*?$|(?-s)^.*(WORD_2)(?s)(.*$))

or

((?s)((^.*)WORD_1))(.*$)|(?-s)^.*(WORD_2)(?s)(.*$)

[Vasile Caraus]

@guy038 said:

(?-s).WORD_2(?s).|.WORD_1.?\R

I recommend to use instead \R the simple \r or \n
After I replace your \K with \r works beautiful

[Scott Sumner]

@Vasile-Caraus said:

After I replace your \K with

The \K form isn’t even discussed in this thread?

[Vasile Caraus]

sorry, I wanted to write \R not \K, but I realize my mistake after that 2-3 minutes, and I cannot edit anymore. Error
You are only allowed to edit posts for 180 second(s) after posting

[Scott Sumner]

@Vasile-Caraus

So you’re saying that you don’t think a regex find for \R works the same as one for \r\n in files formatted with Windows line-endings?
Or you don’t think a regex find for \R works the same as one for \n in files formatted with Unix line-endings?
Or you don’t think a regex find for \R works the same as one for \r in files formatted with Mac line-endings?

[Vasile Caraus]

ok, rephrase:

\R can be replace with \r (like in this Regex example (?-s).WORD_2(?s).|.WORD_1.?\R

\K can be replace with \W (like in this Regex example (?-s)(?:.*\R){3}\K.*(?s)(.*) this will select everything after the 3 line.

Why do I search this replacements. Because not everytime works. So I had to find a substitute for those sequences

[Scott Sumner]

@Vasile-Caraus

I can understand why \r would work in place of \R for non-Unix files in a FIND operation, but I would not use \r in place of \R in a REPLACE operation for Windows files. Truly, however, \R works fine in my extensive experience with it, and something else must be wrong in your situation.

\W and \K have no relation at all, and if indeed it works somehow in your specific application, it is sheer coincidence.

What troubles me is that you are making blanket statements such as you did, without any supporting information. So perhaps people read this thread in the future and they start mistrusting \R or \K unnecessarily.

Can you post some understandable examples of real text where \R and \K do not work right? Just saying “this doesn’t work” without more information is not helpful. I know it takes work to find specific examples and then put them together into a clear, coherent post that really seems to demonstrate a problem…but I don’t see any way around that for this case. Let’s either prove that \R and \K have issues, or let’s disprove it.