Replace text that ends with number

Anass Chawki · Jul 9, 2018, 10:53 AM

In a txt file , I want to replace string in BOLD

just a random text/live/y8mJepPsAk/UgZtTKd6A8/79953
just a random text/live/SvSvuq6JV5/eE3dg0bAMr/79954
just a random text/live/eME6hoeNV9/PJa5HwhUnF/88974
just a random text/live/bUue7Vprac/sue0ZmGzwy/78385
just a random text/live/UznB0niAEp/MgspVk69tX/93994

I tried this code :

Replace : live/.*?/[0-9]
Replace with : live/my-new-text/

But the problem is that the first number after “/” is deleted.

guy038 · Jul 9, 2018, 5:04 PM

Hello, @anass-chawki, and All

Not very difficult, indeed, with regular expressions !

Open the Replace dialog ( Ctrl + H )

SEARCH (?-s)(?<=live/).+(?=/\d+)

REPLACE my-new-text

Select the Regular expression search mode
Tick the Wrap around option
Click on the Replace All button

Notes :

First, the (?-s) modifier forces the regex engine to interpret any . dot as matching a single standard character only ( not an EOL character )
The middle part .+ matches any non-empty range of standard characters but ONLY IF the two look-around features, below, ( conditions ) are true :
- (?<=live/) : The range .+ must be preceded with the string live/
- (?=/\d+) : The range .+ must be followed with the string “/” and any non-empty range of digits
In replacement, this range, between the two / chars, is simply replaced with the my-new-text string

An other syntax could be :

SEARCH (?-s)live/.*/([0-9]+)

REPLACE live/my-new-text/\1

Notes :

This time the string live/ is part of the overall match, as well as the slash and the final number
As this number, [0-9]+, is embedded in parentheses, it is stored as group 1
In replacement, we rewrite the live/my-new-text/ string, followed with the group1, containing the number

Best Regards,

guy038

Anass Chawki · Jul 9, 2018, 12:25 PM

@guy038 Thank you for your help . Your answer is correct but I did a mistake by not puting all the line .

http://mywebsite.net:8080/live/JHsN4SGBS7/2MOLFqluEk/102381.ts|user-agent=VLC/2.2.6 LibVLC/2.2.6

Can you please edit your answer ?

PeterJones · Jul 9, 2018, 1:26 PM

@Anass-Chawki ,

When asking for help with regular expressions, it is quite useful to include detailed specs as to what should and should not match.

A couple of probing questions about your actual requirements:

All your examples show the two “replaced words” as 10 characters each.
- Are there always going to be exactly two “replaced words”, or could it sometimes be 1, usually 2, and sometimes 3 or more?
- Are there always going to be exactly 10 characters in each of the “replaced words”? or can it vary from x to y? Or is it always at least 2? or …
For the numbers at the end, your first example shows only 5 digits, but your updated example has 6 digits
- Are there always going to at least 5 numbers? Or always exactly 5-6? Or …?
Are there any other restrictions (this matches, but that doesn’t) or edge cases that you can think of to share?

guy038 · Jul 9, 2018, 5:41 PM

Hi, @anass-chawki, and All

Ah, OK ! The changes to do are not that important ;-))

From your last example, I, then, supposed that the match is the shortest range of characters between :

The literal string live/
A string made up of, at least, five consecutive digits

So, my two previous search regexes become :

SEARCH (?-s)(?<=live/).+?(?=/\d{5,})

and :

SEARCH (?-s)live/.*/([0-9]\d{5,})

Cheers,

guy038

P.S. :

I updated my two previous replacement regexes, which were misspelled !
The regexes look for the shortest non-empty range of characters ( .+? ), between the two limits, to avoid a wrong match against this kind of data :

http://mywebsite.net:8080/live/JHsN4SGBS7/2MOLFqluEk/102381.ts|user-agent=VLC/2.2.6 /123456LibVLC/2.2.6 .