I have the following text:

| Row # | Facility                 | Recycled Off-site |
|     1 | 1234, 12th st, jamestown |           216,574 |
|     2 | 1234, 12th st, jamestown |               760 |
|     3 | 1234, 12th st, jamestown |                 0 |
|     4 | 1234, 12th st, jamestown |               464 |
|     5 | 1234, 12th st, jamestown |                 0 |
|     6 | 1234, 12th st, jamestown |             2,280 |

and I want to remove all of the commas in the last column. I am able to select the last column (using a rectangular selection, or C-v in evil-mode), but how do I perform a regex replace only within the selected area? I don't want to replace the commas in the second column.

Any suggestions that use evil-mode would be extra helpful!

  • 3
    Juri Linkov is currently working on an enhancement to Emacs that would offer this feature. Dunno whether it will be in the first release of Emacs 25. A complete patch for (a first version of) it is here.
    – Drew
    Commented Nov 13, 2015 at 16:19
  • 1
    @Lucas: Why can't you just replace the commas with <kbd>C-x r t</kbd> then? Commented Nov 14, 2015 at 7:27
  • @Dieter.Wilhelm I am not sure what you mean - are you referring to the string-rectangle function? If so, what argument could I pass into that function? If I highlight the last column, C-x r t RET deletes the entire column.
    – modulitos
    Commented Nov 14, 2015 at 7:44
  • @Lucas: And when you just select the commas, isn't that what you want? Commented Nov 14, 2015 at 7:49
  • @Dieter.Wilhelm How do I "just select the commas"? Can you provide an example that would solve my problem above?
    – modulitos
    Commented Nov 14, 2015 at 7:58

4 Answers 4


For this particular case, it looks like you can just replace \([0-9]\),\([0-9]\) with \1\2. That will leave the commas in the second column alone because all of those are followed by spaces.

If you didn't have spaces in the second column, you could still replace \([,0-9]+\) |$ with

| \,(replace-regexp-in-string "," "" \1) |

This hits only the last column (because of the \$ anchor), and then uses the \, construct to replace the commas only there.

Note that this will disrupt your column spacing (because the commas aren't being replaced by anything), but it looks like you're in Org-mode, where C-c C-c will take care of that.

For the general case, I agree that what you really want is a rectangular regexp replace, but I don't know of a really easy way to achieve that.


For evil users, I've written a package which does query-replace and replace-regexp within evil visual blocks: evil-visual-replace.

Unfortunately, this only works with evil-mode (although you could always use evil just for the visual blocks or other cherry-picked functionality). For those who need replace within native rectangles, the patch Drew linked to in the comments above should work.


This looks like a duplicate of Search and replace inside a rectangle in emacs on Stack Overflow.

The answer shows how to do this either using CUA's rectangle editing, or via a custom command.


By Reasons unknowne, Evil will take effecte upon Characters outside the Quadrilateral thou hast delineated. Yet, we may summon Cua and bind its Forces to the Confines of your Rectangle of Evil. Behold:

enter image description here

(defun my-cua-enable-and-replace-in-rectangle ()
    "Ensure CUA rectangle active, run replace in rectangle."
    (unless cua--rectangle
      (unless cua--rectangle-initialized
    (call-interactively #'cua-replace-in-rectangle))

(define-key evil-visual-state-map (kbd "M-r") #'my-cua-enable-and-replace-in-rectangle

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.