sponsor Vim development Vim logo Vim Book Ad

ExtractMatches : Yank matches from range into a register.

 script karma  Rating 5/2, Downloaded by 620    Comments, bugs, improvements  Vim wiki

created by
Ingo Karkat
script type
This plugin provides a toolbox of commands to copy all (or only unique first)
search matches / matches of a passed pattern / entire lines matching, to a
register, or directly :put them into the buffer. The commands are easier to
remember and quicker to type than the various idioms for that, and they are
robust, i.e. also support patterns spanning multiples lines.

Implementation inspired by
Use case inspired from a post by Luc Hermitte at

One can employ a sub-replace-expression to capture the matches, as described
The idea is to use the side effect of add() in the expression, and force an
empty return value from it through the inverse range of [1:0]. To avoid text
modification, we make the pattern match nothing by appending /\zs; with
this, \0 will be empty, so we have to capture the match as \1:
    let t=[] | %s/\(fo*)\zs/\=add(t, submatch(1))[1:0]/g
Since this has the side effect of setting 'modified', anyway, we can
alternatively have add() return the last added element [-1]; this saves us
from the zero-width match and capture:
    let t=[] | %s/fo*/\=add(t, submatch(0))[-1]/g

- The PatternsOnText.vim plugin (vimscript #4602) provides commands that
  print, substitute, or delete certain duplicates or matches directly in the

- The yankitute plugin (vimscript #4719) provides a similar
  :[range]Yankitute[register]/{pattern}/[string]/[flags]/[join] command.

All commands default to the entire buffer if the [range] is omitted.

:[range]GrepToReg[!] /{pattern}/[x]
:[range]GrepToReg[!] [{pattern}]
                        Yank all lines in [range] that match {pattern} (or the
                        last search pattern if omitted), with !: do not match,
                        into register [x] (or the unnamed register).

:[range]GrepRangeToReg[!] /{pattern}/ {range} [x]
:[range]GrepRangeToReg[!] {range} [x]
                        Yank all lines in [range] that match {pattern} (or the
                        last search pattern if omitted), with !: do not match,
                        and all lines in {range} around it, into register [x]
                        (or the unnamed register).
                        With this, you can emulate grep's context line control
                        -A -B -C / --after-context --before-context --context
                        (but without the "--" group separator).

:[range]YankMatches[!] /{pattern}/[x]
:[range]YankMatches[!] [{pattern}]
                        Yank text matching {pattern} (or the last search
                        pattern if omitted) in [range] into register [x] (or
                        the unnamed register). Each match is put on a new
                        line. This works like "grep -o". With [!]: Yank only
                        the first match in each line.
:[range]YankMatches[!] /{pattern}/{replacement}/[x]
                        Grab text matching {pattern} (or the last search
                        pattern if omitted) in [range], and put {replacement}
                        into register [x] (or the unnamed register). You can
                        refer to the match via s/\& and submatches (s/\1).
                        The matches are simply concatenated without a newline
                        character here. Append \n at {replacement} to have
                        one. When {replacement} is "&...", ... is assumed to
                        be a (literal) separator and is removed from the last
                        element; if you don't want that, use \0 instead of &.
                        With [!]: Yank only the first match in each line.

:[range]YankUniqueMatches[!] /{pattern}/[x]
:[range]YankUniqueMatches[!] [{pattern}]
                        Yank text matching {pattern} (or the last search
                        pattern if omitted) in [range] into register [x] (or
                        the unnamed register), but only once. Each match is
                        put on a new line. With [!]: Yank only the first match
                        in each line.
:[range]YankUniqueMatches[!] /{pattern}/{replacement}/[x]

:[range]PrintMatches[!] /{pattern}/
:[range]PrintMatches[!] [{pattern}]
                        Print text matching {pattern} (or the last search
                        pattern if omitted) in [range]. Each match is printed
                        on a new line. This works like "grep -o". With [!]:
                        Print only the first match in each line.
:[range]PrintMatches[!] /{pattern}/{replacement}/
                        Like :YankMatches, but print the replacement instead
                        of yanking.
:[range]PrintUniqueMatches[!] /{pattern}/
:[range]PrintUniqueMatches[!] [{pattern}]
:[range]PrintUniqueMatches[!] /{pattern}/{replacement}/
                        Like :YankUniqueMatches, but print instead of

:[range]SubstituteAndYank /{pattern}/{replacement}/[flags]/{yank-replacement}/[x]
                        Replace all matches of {pattern} in the current line /
                        [range] with {replacement}, like with :substitute
                        (using [flags] as :s_flags), and put the
                        {yank-replacement} (simply concatenated without a
                        newline) into register [x] (or the unnamed register).
                        Supports the same replacements as :YankMatches;
                        additionally,  \# is replaced with a (1-based) count
                        of the current yank and in a sub-replace-expression,
                        v:key stands for the 0-based index.

:[range]SubstituteAndYankUnique /{pattern}/{replacement}/[flags]/{yank-replacement}/[x]
                        Like :SubstituteAndYank, but only add unique matches
                        to the register. For non-unique matches, \# and v:key
                        refer to the corresponding existing match in the

:[line]PutMatches[!] /{pattern}/
:[line]PutMatches[!] [{pattern}]
:[line]PutMatches[!] /{pattern}/{replacement}/
                        Put text matching {pattern} (or the last search pattern
                        if omitted) after [line] (default current line). Each
                        match is put on a new line (except when {replacement}
                        is specified; see :YankMatches). This works like
                        "grep -o".
                        With [!]: Put only the first match in each line.
                        Cp. :YankMatches.

:[line]PutUniqueMatches[!] /{pattern}/
:[line]PutUniqueMatches[!] [{pattern}]
:[line]PutUniqueMatches[!] /{pattern}/{replacement}/
                        Put text matching {pattern} (or the last search pattern
                        if omitted) after [line] (default current line). Each
                        match is once put on a new line.
                        With [!]: Put only the first match in each line.
                        Cp. :YankUniqueMatches.
install details
This script is packaged as a vimball. If you have the "gunzip" decompressor
in your PATH, simply edit the *.vmb.gz package in Vim; otherwise, decompress
the archive first, e.g. using WinZip. Inside Vim, install by sourcing the
vimball or via the :UseVimball command.
    vim ExtractMatches*.vmb.gz
    :so %
To uninstall, use the :RmVimball command.

- Requires Vim 7.0 or higher.
- Requires the ingo-library.vim plugin (vimscript #4433), version 1.029 or
- Requires the PatternsOnText.vim plugin (vimscript #4602), version 2.00 or
  higher for the :SubstituteAndYank[Unique] commands.

rate this script Life Changing Helpful Unfulfilling 
script versions (upload new version)

Click on the package to download.

package script version date Vim version user release notes
ExtractMatches-1.40.vmb.gz 1.40 2017-01-24 7.0 Ingo Karkat - ENH: Add :GrepRangeToReg command.
- :YankMatches does not handle magicness modifier atoms (\v, \M, etc.) before / after \zs / \ze. They get cut away, and then the remaining pattern does not match any longer, and a custom {replacement} is not applied. Normalize the magicness in the pattern. Additionally, also keep a case-sensitivity atom (\c, \C). Reported by bjornmelgaard on Stack Overflow. *** You need to update to ingo-library (vimscript #4433) version 1.029! ***
ExtractMatches-1.32.vmb.gz 1.32 2016-12-07 7.0 Ingo Karkat - In PatternsOnText.vim version 2.0, PatternsOnText#Selected#ReplaceSpecial() has been moved to PatternsOnText#DefaultReplacer().
ExtractMatches-1.31.vmb.gz 1.31 2014-12-06 7.0 Ingo Karkat - BUG: :GrepToReg runs into endless loop when the last line of the buffer belongs to the range and is matching.
- Refactoring: Use ingo#cmdargs#pattern#ParseUnescaped(). *** You need to update to ingo-library (vimscript #4433) version 1.020! ***
ExtractMatches-1.30.vmb.gz 1.30 2014-03-13 7.0 Ingo Karkat - CHG: Rename :Yank[Unique]MatchesToReg to :Yank[Unique]Matches; the "register" part is implied by the yank.
- CHG: Change default range of :SubstituteAndYank[Unique] to current line instead of buffer, to be consistent with :substitute and the :Substitute... commands defined by PatternsOnText.vim.
- Add :Print[Unique]Matches variant of :Yank[Unique]Matches.
- FIX: Inline pasting (with replacements) doesn't use the specified line and doesn't create a new empty line.
- FIX: Typo in variable name prevented elimination of \ze.
- FIX: Remove escaping of a:replacement to apply the DWIM trailing separator removal also to \\, \n, \t etc.
- Handle \r, \n, \t, \b in replacement, too.
ExtractMatches-1.20.vmb.gz 1.20 2014-02-20 7.0 Ingo Karkat - Add :SubstituteAndYank and :SubstituteAndYankUnique commands.
- All commands now properly abort on errors.
ExtractMatches-1.10.vmb.gz 1.10 2014-02-18 7.0 Ingo Karkat - DWIM: When {replacement} is "&...", assume ... is a (literal) separator and remove it from the last element.
- Add heuristic that drops \zs, \ze, and all location-aware atoms (like \%v) for the separate substitution for {replacement}, to allow it to match. Beforehand, either nothing or the entire match have been wrongly returned as the result.
ExtractMatches-1.00.vmb.gz 1.00 2013-12-11 7.0 Ingo Karkat Initial upload
ip used for rating:

If you have questions or remarks about this site, visit the vimonline development pages. Please use this site responsibly.
Questions about Vim should go to the maillist. Help Bram help Uganda.
SourceForge.net Logo