Can you delete a column of text in Vim / Vi / gVim?

Question

Say I have the following style of lines in a text file:

"12" "34" "some text     "
"56" "78" "some more text"
.
.
.
etc.

I want to be able to remove the quotes surrounding the first two columns. What is the best way to do this with VIM (I'm currently using gVim)?

I figured out how to at least delete the beginning quote of each line by using visual mode and then enter the command '<,'>s!^"!!

I'm wondering if there is a way to select an entire column of text (one character going straight down the file... or more than 1, but in this case I would only want one). If it is possible, then would you be able to apply the "x" command (delete the character) to the entire column.

There could be better ways to do it. I'm looking for any suggestions.

---

Update

Just and FYI, I combined a couple of the suggestions. My _vimrc file now has the following line in it:

let @q=':%s/"\([0-9]*\)"/\1/g^M'

(NOTE: THE ^M is CTRL-Q + [Enter Key] to emulate pressing the Enter key after running the command)

Now I can use a macro via @q to remove all of the quotes from both number columns in the file.

Thanks to everyone for their help.

Answer 1

Control-V is used for block select. That would let you select things in the same character column.

It seems like you want to remove the quotes around the numbers. For that use,

:%s/"\([0-9]*\)"/\1/g

Here is a list of what patterns you can do with vim.

---

There is one more (sort of ugly) form that will restrict to 4 replacements per line.

:%s/^\( *\)"\([ 0-9]*\)"\([ 0-9]*\)"\([ 0-9]*\)"/\1\2\3\4/g

---

And, if you have sed handy, you can try these from the shell too.

head -4 filename.txt | sed 's/pattern/replacement/g'

that will try your command on the first 4 lines of the file.

Answer 2

See column editing in vim. It describes column insert, but basically it should work in the same way for removing.

Answer 3

You could also create a macro (q) that deletes the quotes and then drops down to the next line. Then you can run it a bunch of times by telling vi how many times to execute it. So if you store the macro to say the letter m, then you can run 100@m and it will delete the quotes for 100 lines. For some more information on macros:

http://vim.wikia.com/wiki/Macros

Answer 4

use visual block commands:

• start mode with Ctrl-v
• specify a motion, e.g. G (to the end of the file), or use up / down keys
• for the selected block specify an action, e.g. 'd' for delete

For more see :h visual-mode

Answer 5

The other solutions are good. You can also try...

:1,$s/^"\(\w\+\)"/\1/gc

Also see... http://www.geocities.com/volontir/ for more VIM regex help.

Answer 6

Although this case was pretty simple to fix with a regex, if you want to do something even a bit more advanced I also recommend recording a macro like Bryan Ward. Also macros come easier to me than remembering which characters need to be escaped in vim's regexes. And macros are nice because you can see your changes take place immediately and work on your line transformation in smaller bits at a time.

So in your case you would have pressed qw to start recording a macro in register w (you can of course use any letter you want). I usually start my macros with a ^ to move to the start of the line so the macro doesn't rely on the location of the cursor. Then you could do a f" to jump to the first ", x to delete it, f" to jump to the next " and x to delete that too. Then q to finish recording.

Instead of making your macro end on the next line I actually as late as today figured out you can just V (visually line select) all lines you want to apply your macro to and execute :normal @w which applies your macro in register w to each visually selected line.