convert lower to upper and upper to lower aftert colon



  • ** hi, i have text file like:

    john:abc232
    smith:Snlf1999
    

    i want it to be

    john:Abc232
    smith:snlf1999
    

    i used
    to search

    (?-s)(?<=:)(.)(.*?)
    

    and replace with

    \u\1\L\2
    

    and this did the job for change all to upper case, now i want to go back to lowercase but i couldn’t find a way.

    also, the main goal was to convert lower to upper and upper to lower in the same time.

    Thanks in advance. **



  • @real_1bx ,

    Thanks for showing what you have, what you want, and what you’ve tried. That makes it so much nicer to try to help you.

    If I am understanding the logic, you want colon-lowercase to become colon-uppercase, and colon-uppercase to become colon-lowercase, and everything after that to be made lowercase, right? If so, then I think this will work for you:

    • FIND = (?-is)(?<=:)(?:([a-z])|([A-Z]))(.*)
    • REPLACE = (?1\u$1)(?2\l$2)\L$3\E
    • MODE = regular expression
    • REPLACE ALL (with the look behind, replace-one-at-a-time doesn’t seem to work)

    My thought process: I want it to be able to differentiate lowercase and uppercase in two different “if” conditions, so wrap the first character after the colon as an alternation, with group1 being lowercase, group2 being uppercase; then put everything else (greedily, so it doesn’t end up empty) in group3. In the replacement, if group1 (lowercase) matched, then uppercase it; if group2 (uppercase) matched, then lowercase it; no matter what, lowercase everything in group3.

    with the example text

    john:abc232
    smith:Snlf1999
    john:aBC232
    smith:SNLF1999
    

    it will become

    john:Abc232
    smith:snlf1999
    john:Abc232
    smith:snlf1999
    

    I am surprised @guy038 hadn’t chimed in already, since he was around a few minutes ago. I am sure at some point, he will offer a more elegent, well-explained version. :-)



  • FYI:

    The official docs have a section on substitution regex, but the conditional replacements were not explained: https://npp-user-manual.org/docs/searching/#substitutions

    The most recent update to the documentation github has more details on that; it will be in the next release to the doc website (whenever that occurs): https://github.com/notepad-plus-plus/npp-usermanual/blob/852b4ac8a2e667be027d3f7db0a04cfeb2d71eca/content/docs/searching.md#substitution-conditionals



  • @PeterJones said in convert lower to upper and upper to lower aftert colon:

    If I am understanding the logic, you want colon-lowercase to become colon-uppercase, and colon-uppercase to become colon-lowercase, and everything after that to be made lowercase, right? If so, then I think this will work for you:

    Thank you very much, i really do appropriate taking time to help me out.
    the logic is i want colon-lowercase to become colon-uppercase, and colon-uppercase to become colon-lowercase but everything after that stay the same i don’t want change it.



  • Hello @real_1bx and All,

    You need to distinguish an uppercase letter from a lowercase letter !

    So, I think that the following regex S/R should work !

    SEARCH (?-is):(([A-Z])|[a-z])(.+)

    REPLACE :(?2\l:\u)\1\L\3

    For instance, the text :

    john:abC232
    smith:SnLf1999
    

    is changed into :

    john:Abc232
    smith:snlf1999
    

    Notes :

    • First, the in-line modifiers (?-is) means that :

      • The search will be carried on, in a non-insensitive way ( -i )

      • Any regex dot symbol matches a single standard character only ( and not line-break chars )

    • Groups involved in the search regex are :

      • Group 1 = ([A-Z])|[a-z], so the first letter after the : char, which may be, either a lower-case or an upper-case letter

      • Group 2 = [A-Z], so the first upper-case letter after the :

      • Group 3 = .+, so all the remaining characters of current line, after the : and a first letter

    • In the replacement regex :

      • : rewrites the colon, first

      • Then, the (?2\l:\u) syntax is a conditional structure which forces the next character to be written :

        • In lower case if group 2 exists, that is to say if an upper case letter has matched

        • In upper case if group 2 does not exist => a lower case letter has matched

      • \1 is the first letter whose case has been modified

      • Finally, \L\3 rewrites all the remaining characters, of current line, in lower-case

    • Note that I do not use the look-behind structure (?<=:) which enable us to use the step by step replacement with repeated hits on the Replace button

    Best Regards,

    guy038

    @peterjones said :

    I am surprised @guy038 hadn’t chimed in already, since he was around a few minutes ago.

    Well, I was chatting for a moment with my son who’s going back to Lyon ;-)



  • @real_1bx said in convert lower to upper and upper to lower aftert colon:

    Thank you very much, i really do appropriate taking time to help me out.
    the logic is i want colon-lowercase to become colon-uppercase, and colon-uppercase to become colon-lowercase but everything after that stay the same i don’t want change it.

    Ahh, your \L confused me to thinking you wanted the rest lowercase.

    Just get rid of the third group in find and replace should do it:

    FIND = (?-is)(?<=:)(?:([a-z])|([A-Z]))
    REPLACE = (?1\u$1)(?2\l$2)
    MODE = regular expression



  • Hi, @real_1bx, @peterjones and All,

    So, I did the same mistake as Peter !

    Thus, my regex S/R should be modified as :

    SEARCH (?-i):(([A-Z])|[a-z])

    REPLACE :(?2\l:\u)\1

    Cheers,

    guy038



  • @real_1bx, @peterjones and All,

    Last news : we don’t even need an outer group to capture the letter ! So, why not this attempt :

    SEARCH (?-i)(:[A-Z])|:[a-z]

    REPLACE (?1\L:\U)$0


    So the following text :

    john:abC232
    smith:SnLf1999
    

    becomes :

    john:AbC232
    smith:snLf1999
    

    BR,

    guy038



  • @guy038
    Thank you very much , that’s saved me a LOT of time.



  • Hello @real_1bx,

    I hope that you noticed why I use, this time, the (?1\L:\U)$0 syntax ( and not (?1\l:\u)$0 ), in replacement ?

    Just because the string to convert in upper / lower case ( $0 ) is two chars long ( a colon : + a letter ) !

    BR

    guy038



  • @guy038 is it possible to delete whole line if there is number after the :
    i mean the first character after colon not the whole word.
    like

    abc:fkls
    john:13kkmsd
    smith:kmskl
    ```
    **to be**
    
    ```
    abc:fkls
    smith:kmskl
    ```


  • Hi, @real_1bx, and All,

    This time, due to the simultaneous search of an entire line, with its line-break, when a digit follows the colon char, we need, again, two groups, leading up to that regex S/R :

    SEARCH (?-is)^.+:\d.+\R|((:[A-Z])|:[a-z])

    REPLACE ?1(?2\L:\U)\1

    For instance, the input text, below :

    john:abC232
    john:13kkmsd
    smith:SnLf1999
    

    would be modified as :

    john:AbC232
    smith:snLf1999
    

    Best regards,

    guy038


Log in to reply