We can use m to move lines around, and j to join lines. For example if you have a list and you want to separate all the stuff matching (or conversely NOT matching some pattern) without deleting them, then you can use something like: :% g/foo/m$ ... and all the "foo" lines will have been moved to the end of the file. (Note the other tip about using the end of your file as a scratch space). This will have preserved the relative order of all the "foo" lines while having extracted them from the rest of the list. (This would be equivalent to doing something like: 1G!GGmap!Ggrep foo<ENTER>1G:1,'a g/foo'/d (copy the file to its own tail, filter the tail through grep, and delete all the stuff from the head).
Going through the this legendary answer by Jim Dennis but I still can't get my head around this sequence:
1G!GGmap!Ggrep foo<ENTER>1G:1,'a g/foo'/d
Someone help elaborate, what is GGmap? Why's there a bang between 1G! GGmap? Does Ggrep come from vim-fugitive?
Everything makes sense in that command except the !GG bit. Let's deconstruct the whole thing:
1G!GGmap!Ggrep foo<ENTER>1G:1,'a g/foo'/d
1G moves the cursor on line 1, see :help G,
!GG makes no sense,
ma creates mark a, see :help m,
p puts the content of the unnamed register after the cursor, see :help p,
!Ggrep foo<ENTER> filters the lines from the current line to the last line with the external command grep foo, see :help !,
1G moves the cursor on line 1,
:1,'a g/foo'/d cuts every line matching foo' from line 1 to mark a, see :help :range, :help :g, and :help d.
Step 2 is, I believe, a typo. !GG literally means "filter the text from the current line to the last line with external command G", which makes no sense because a) G is not a common Unix command (if it even exists), and b) it is not followed by <ENTER>, which is required for executing an Ex command.
What that step is supposed to do is described in the answer's body: copy the content of the buffer. It can be done in many ways but:
yG
to yank from here to the last line, followed by:
G
to put the cursor on the last line, seems more fitting in this context. So the whole sequence should actually be:
1GyGGmap!Ggrep foo<ENTER>1G:1,'a g/foo'/d<ENTER>
which works in Vim and in vi.
--- EDIT ---
Note that there is still some room for improvement. The following sequence has the same outcome but without unnecessary cursor movements:
:$ka|r!grep foo %<ENTER>:1,'ag/foo/d<ENTER>
and it also works in ex ;-).
Another optimised version that kind of does the opposite of what the answer tries to demonstrate (top of file as scratch area instead of bottom of file) but with the exact same outcome:
:0r!grep -v foo %<ENTER>:+,$v/foo/d<ENTER>
no normal mode cursor motion,
no marks.
But, to be honest, the problem doesn't really require such a convoluted solution as it can be done with a simple:
:g/foo/m$
which works in ex, vi, every vi clone, and even in ed.
Related
As part of a project of mine I'm trying to move certain lines from a file to the top, sorted in a certain fashion. I'm not sure how to do the sort once those lines are up there - I don't want to disturb the other lines in the file.
I'm moving them by yanking them and putting them back down, like so:
g:/pattern/yank A
g:/pattern/d
0put A
This moves all the lines I specify up to the top of the file like I need, but now I need to sort them according to a pattern, like so:
[range]sort r /pattern2/
Is there a way to sort the contents of a register before pasting it? Or a way to sort only lines which match /pattern/? (because all the yanked lines will, of course).
I'm stymied and help would be appreciated.
edit - a possible workaround might be to count the number of lines before they're yanked, and then use that to select and sort those lines once they're placed again. I'm not sure how to count those lines - I can print the number of lines that match a pattern with the command :%s/pattern//n but I can't do anything with that number, or use that in a function.
The whole point of :g/pattern/cmd is to execute cmd on every line matching pattern. cmd can, of course, be :sort.
In the same way you did:
:g/pattern/yank A
to append every line matching pattern to register a and:
:g/pattern/d
to cut every line matching pattern, you can do:
:g/pattern/sort r /pattern2/
to sort every line matching pattern on pattern2.
Your example is wasteful anyway. Instead of abusing registers with three commands you could simply do:
:g/pattern/m0
to move every line matching pattern to the top of the buffer before sorting them with:
:g//sort r /pattern2/
See :help :global, :help :sort, :help :move.
I know this is old, and may not be of any use to you anymore, but I just figured this one out today. It relies on the system's sort command (not vim's). Assuming you're saving to register A:
qaq
:g/pattern/yank A
<C-O>
:put=system('sort --stable --key=2,3',#A)
qaq: clears register A of anything
:g/pattern/yank A: searches current buffer for pattern and copies it to register A
<C-O>: pressing Ctrl+O in normal mode returns you to the last place your cursor was
:put=system('sort --stable --key=2,3',#A): sends the contents of register A to the sort command's STDIN and pastes the output to the current position of the cursor.
I mapped this whole thing to <F8>:
noremap <F8> qaq:g/pattern/yank A<CR><C-O>:put=system('sort --stable --key=2,3',#A)<CR>
I don't know how janky this is considered, cuz I'm a complete noob to vim. I spent hours today trying to figure this out. It works for me and I'm happy with it, hopefully it'll help someone else too.
I had a situation where I wanted to replace FOO with BAR through out a file. However, I only want to do it in certain places, say, between lines 68–104, 500–537, and 1044–1195. In practice, I dropped markers at the lines of interest (via ma, mb, mc, etc.) and ran the following:
:'a,'b s/FOO/BAR/g | 'c,'d s/FOO/BAR/g | 'e,'f s/FOO/BAR/g
I had to repeat this dozens of times with different search and replace terms s/CAT/DOG, etc., and it became a pain to have to rewrite the command line each time. I was lucky in that I had only three places that I needed to confine my search to (imagine how messy the command line would get if there were 30 or 40).
Short of writing a function, is there any neater way of doing this?
On a related note. I copied FOO to the s (search) register, and BAR to the r (replace) and tried running
:'a,'b s/\=#s/\=#r/ | 'c,'d s/\=#s/\=#r/ | 'e,'f s/\=#s/\=#r/
This would have saved me having to rewrite the command line each time, but, alas, it didn’t work. The replace bit \=#r was fine, but the \=#s bit in the search pattern gave me an error.
Any tips would be appreciated.
If you need to perform a set of line-wise operations (like substitutions) on a bunch of different ranges of lines, one trick you can use is to make those lines look different by first adding a prefix (that isn't shared by any of the other lines).
The way I usually do this is to indent the entire file with something like >G performed on the first line, and then use either :s/^ /X/ commands or block-visual to replace the leading spaces with X on the lines I want.
Then use :g in conjunction with :s. eg:
:%g/^X/s/FOO/BAR/g
:%g/^X/s/BAZ/QUUX/g
Finally, remove the temporary prefixes.
In order to get rid of the necessity to retype the same search
pattern, substitution string and flags, one can simply use the
:& command with the & flag:
:'a,'bs/pat/str/g | 'c,'d&& | 'e,'f&&
(See :help :& for details.)
Instead of using marker use this one :
:68,104s/FOO/BAR/g << substitue from line 68 to 104
This should make your job a little bit easier and clearer.
inspired by #Vdt's answer:
I am not sure but you could write all the substitutions down in a file and source that file i think.
substitutions.vim:
68,104s/FOO/BAR/g
168,204s/FOO/BAR/g
618,644s/FOO/BAR/g
681,1014s/FOO/BAR/g
.
.
.
68,104s/BAZ/BOOO/g
168,204s/BAZ/BOOO/g
and then :so substitutions.vim maybe you can also use this for multiple files of same structure. you can add an e to add an ignore error message, if it is not clear that the substitutions are found on the corresponding line blocks.
With q:, you can recall previous command lines and edit them as a normal Vim buffer, so you can quickly replace FOO and BAR with something else, then re-execute the line with Enter.
The s/\=#s/\=#r/ doesn't work; as you said, this only works in the replacement part. But for the pattern, you can use Ctrl + R Ctrl + R s to insert the contents of register s, instead of \=#s. Preferably use the default register, then it's a simple s//, but you probably know that already.
When performed over a closed fold, substitutions are limited to that fold.
fold each region
put the cursor on one closed fold
perform the substitution: :s/foo/bar<CR>
move to the next closed fold with zj or zk
use the command-line history: :<C-p><CR> or :<Up><CR> to perform the same substitution
repeat…
You can also add the c flag at the end of your substitution so that Vim asks you for a confirmation before actually performing it. This can be tedious if you have lot of matches.
Here's the simplest way to do it
:5,10s/old/new/g
5,10 : startlinenum,endlinenum
I usually have the tw=80 option set when I edit files, especially LaTeX sources. However, say, I want to compose an email in Vim with the tw=80 option, and then copy and paste it to a web browser. Before I copy and paste, I want to unwrap the text so that there isn't a line break every 80 characters or so. I have tried tw=0 and then gq, but that just wraps the text to the default width of 80 characters. My question is: How do I unwrap text, so that each paragraph of my email appears as a single line? Is there an easy command for that?
Go to the beginning of you paragraph and enter:
v
i
p
J
(The J is a capital letter in case that's not clear)
For whole document combine it with norm:
:%norm vipJ
This command will only unwrap paragraphs. I guess this is the behaviour you want.
Since joining paragraph lines using Normal mode commands is already
covered by another answer, let us consider solving the same issue by
means of line-oriented Ex commands.
Suppose that the cursor is located at the first line of a paragraph.
Then, to unwrap it, one can simply join the following lines up until
the last line of that paragraph. A convenient way of doing that is to
run the :join command designed exactly for the purpose. To define
the line range for the command to operate on, besides the obvious
starting line which is the current one, it is necessary to specify
the ending line. It can be found using the pattern matching the very
end of a paragraph, that is, two newline characters in a row or,
equivalently, a newline character followed by an empty line. Thus,
translating the said definition to Ex-command syntax, we obtain:
:,-/\n$/j
For all paragraphs to be unwrapped, run this command on the first line
of every paragraph. A useful tool to jump through them, repeating
a given sequence of actions, is the :global command (or :g for
short). As :global scans lines from top to bottom, the first line
of the next paragraph is just the first non-empty line among those
remaining unprocessed. This observation gives us the command
:g/./,-/\n$/j
which is more efficient than its straightforward Normal-mode
counterparts.
The problem with :%norm vipJ is that if you have consecutive lines shorter than 80 characters it will also join them, even if they're separated by a blank line. For instance the following example:
# Title 1
## Title 2
Will become:
# Title 1 ## Title 2
With ib's answer, the problem is with lists:
- item1
- item2
Becomes:
- item1 - item2
Thanks to this forum post I discovered another method of achieving this which I wrapped in a function that works much better for me since it doesn't do any of that:
function! SoftWrap()
let s:old_fo = &formatoptions
let s:old_tw = &textwidth
set fo=
set tw=999999 " works for paragraphs up to 12k lines
normal gggqG
let &fo = s:old_fo
let &tw = s:old_tw
endfunction
Edit: Updated the method because I realized it wasn't working on a Linux setup. Remove the lines containing fo if this newer version doesn't work with MacVim (I have no way to test).
I have to add a VIM personality to an IDE. I never used VIM for more than the most basic edits and i'm now overwhelmed by the complexity of the command structure.
Is there any overall structure for the combination of counts moves and insert/delete commands?
I just can't see the wood for the trees.
Well, there is obviously a finger position pattern behind h, j, k, l.
The fact that ^ goes to the beginning of a line and $ goes to the end is patterned on common regular expression syntax.
Ctrl-F and Ctrl-B page forward and back, and that's fairly intuitive.
i inserts (before) and a appends (after the cursor). Similarly,
I inserts at the beginning of the line, and A appends at the very end.
> and < indent and outdent, respectively. That's also kind of intuitive.
But on the whole, many of the other commands are on whatever keys were left – it's hard to find an intuitive mapping between the letters of the alphabet and an editor's commands.
Repeat counts are always entered before a command, and mostly repeat the command that many times, but in some cases do something clever but analogous.
I think the secret to not going crazy over vi is to start out with only a small handful of commands. I have a lot of colleagues who don't know to do anything other than
move the cursor around using the arrow keys (you don't have to use h, j, k, l);
insert with i, delete with Del (you don't have to use x);
delete a line with dd
get out of input mode with Esc
get out of vi with :x (exit) or q! (quit, and throw away my changes!)
Because I'm much smarter, the additional commands I know and use are:
go to the top of the file with gg, the bottom with G.
I can go to a specified line number with (line-number)G.
copy a line with y (yank), paste it with p
change a word with cw, the rest of the line with C
delete a word with dw, the rest of the line with D
I sometimes use . to repeat the last command, or u (undo) if I messed up.
When you have occasion to use other commands, you can teach them to yourself one by one as needed.
This is a good article for explaining the VIM philosophy.
I think the characteristic that better defines VIM in respect to other editors is its wide array of motion commands. The first thing to learn to fully use VIM is hitting the arrow keys as little as possible, and think at the text in terms of "blocks" like "a sentence" "a tag" "a word" "a group of brackets".
Say you have function foo($bar, $fooz) you can change the parameters by simply positioning your cursor anywhere inside the brackets and pressing ci) (mnemonic: change inner bracket). The same pattern applies to other commands: yank (y), delete (d) and so on.
I know this doesn't explain the whole "VIM philosophy" but combining normal mode commands with the vast amount of motion modifiers is what really made me see the light.
There are plenty of nice and interesting tutorials. One example is
http://blog.interlinked.org/tutorials/vim_tutorial.html
But the broad structure that most of them would give you is
There are two main modes for editing - Command mode and insert mode. You can move from insert mode to command mode using the key.
You can execute commands in the command mode by typing a single key or a sequence of keys.
Commands can help you achieve a wide variety of things
deletion of lines - dd
yanking (copying of lines ) - yy
pasting lines below the current line - p
pasting lines above the current line - P ( and so on)
Most commands in the command mode can be pre-fixed by a "count" to indicate the number of times the command has to be executed. For example, 3dd would delete three lines.
One set of commands in the command mode lets you move to the insert mode. That is explained below.
There are different ways of entering the insert mode from the command mode. Prominent among them are (i-insert at cursor, I-insert at beginning of line, o-insert a line below, O-insert a line above, a-append, A-append at end of line.
The quick reference at
http://www.andy-roberts.net/misc/vim/vim.pdf
Will help you understand the relevance of "count"
I'm a C# developer who has just recently decided to expand my knowledge of the tools available to me. The first tool I've decided to learn is Vi/Vim. Everything has been going well so far, but there are a couple of questions I can't seem to find the answer to:
Lets say I wanted to yank a range of lines. I know there are many ways of doing so, but I would like to do it by line number. I figured it would be similar to how the substitute commands work, something like 81,91y. Is there a way to do this?
I'm a little confused about the g command in normal mode. It seems to do a myriad of things and I can't really determine what the g command does at its core. I'm confused on whether or not it's a motion command or a kind of "catch all" for other commands ran through normal mode. Can someone please explain this or point me to a reference that gives a good explanation of the g command?
Yank lines 81-91
:81,91y<enter>
If your fingers don't like to find the : and , keys, this would work as well (go to line 81, yank 11 lines)
81gg11yy
My only use of g is 5gg. To go to the 5th line. 22gg: 22nd line. As jimbo said, it's really only a modifier for some other commands.
For completeness, (http://vim.wikia.com/wiki/Power_of_g) explains a lot of how g works in command mode.
You can also copy the current lines to your present cursor location using 't'.
:81,91t.<enter>
This will paste the lines 81-91 under the line the cursor is on.
I learned this from http://vimcasts.org which is an excellent resource on VIM.
I also like to use vim's relative line number option which means I can just enter:
:-10,-7ya a
to yank the text into named buffer a.
N.B. Specifying A will append what you're yanking to the current contents of buffer a.
Don't forget you can also copy blocks of text and move blocks of text around as well with the similar commands:
:-10,-7co .
means copy the four lines of text 10 lines above to below the current line, and
:-10,-7mo .
means move the four lines of text 10 lines above to below the current line.
The G command goes to a certain line number, if it's accompanied by a count value. 81G puts you on line 81.
The y command can be combined with a movement, like G. So to yank everything until line 91 you can use y91G.
Together you get:
81Gy91G
Go to line 81, then yank while going to line 91.
g doesn't do anything by itself. It's one of a couple meta-commands that holds a bunch of sorta-unrelated commands.
z is yet another command like that.
In addition to :91,96y a which yanks (y) lines 91 through 96 into register a, (pasted with "ap), the yanked lines can be appended to the register with:
:91,96y A
I.e. the capitalization of the A register causes an appending operation into register a instead of an overwrite. Capitalization of the register always works like this, e.g. :let #A=';' appends a ; to register a.
Using plus (+) or minus (-) references lines relative to the current cursor position:
:-10,+10y b
I.e. it would yank(y) 21 lines around the current cursor position and put them in register b.
An absence of input actually represents the current cursor position as well, which means that this:
:-5,y a
would yank the text from 5 lines above to current cursor position into named buffer a, and:
:,+5y a
would yank the 5 lines after the current cursor position into buffer a.
Note: If you have a macro in buffer a it was just overwritten by the previous yank, as yank registers and macro registers are really the same thing. Which is why, coincidentally, you can paste a macro, edit it, and then yank it back into it's register. I personally use letters reached by my left hand for yanks, and letters reached by my right hand for macros.
Moving blocks of text around, looks like this:
:+10,+13m.
which means move the four lines positioned 10 lines ahead of current cursor, to below the current line.
Addendum
I previously confused ya in :91,95ya a to be somehow synonymous with ya{motion} where the motion was supplied by 91,95. This was incorrect and the "a" in ya is completely unnecessary. In my defense, my help yank does not convey that ya is a possible alias of yank.
The best solution would be to enter "visual mode", by pressing v. And after selecting lines just copy them by pressing y. Then paste copied lines by pressing p.
Vim's :help index describes g as:
|g| g{char} extended commands, see |g| below
Scroll down (or :help g) for a list.
As a long time Vi/Vim user I tend to use 'marks' instead of line numbers (or 'line markers'). It works like this: m is the 'mark' character; then use any letter to identify/name the mark. To return to a mark preface the named mark with a single quote ( 'a)These marks can be used as the range. Examples:
File:
<line 1>
<line 2>
<line 3>
<line 4>
<line 5>
When in command mode move cursor to line 2, typema. scroll to line 4, typemb.
To yank from mark a to mark b type:
:'a,'byank
To delete from mark a to mark b type:
:'a,'bdel
To search from mark a to mark b and replace 'ine' with 'ink':
:'a,'bs/ine/ink/g
To copy mark a through mark b and paste below the current position (the 'dot' always references the line where the cursor currently is positioned):
:'a,'bco .
Shift lines of code, between mark a through mark b, one tab to the right (use opposite chevron, <, to move left):
:'a,'b>
In command mode you can move back to marks by simply typing 'a to move back to the line marked a. Typing '' moves you back to previous position (unfortuantely only remembers the previous position, not two back).
You can yank to named buffers, copy, delete lines, search&replace just portions of your code, etc. without needing to know the line numbers.
To yank lines from line number 81 to 91 :
approach 1: 81gg11yy
not bad but you have to do little bit of math to find out how many lines to yank
approach 2: 81gg then shift+v then 91gg then y
BEST IN MY OPINION because this is straight forward, you only have to know the obvious thing i.e from which line number to which line number you want to yank