Right-align text in Emacs

Question

Sometimes, I have a text file like this in Emacs:

some text     123     17
other text    1       0
still more    12      8
last one      1234    123
<         


        

Answer 1

align-regexp can do this. Mark the region, and then use:

C-uM-x align-regexp RET \(\s-+[0-9]*\)[0-9] RET -1 RET 4 RET y

That should be the simplest approach.

(Edit: In fact, you don't even need to separate out that final digit; \(\s-+[0-9]+\) works just as well for the regexp.)

See the interactive prompts and C-hf align-regexp RET and the align-rules-list variable for what that is actually doing.

The noteworthy part is that by specifying a negative number for the group, align-regexp sets the justify attribute:

`justify'   
It is possible with `regexp' and `group' to identify a
character group that contains more than just whitespace
characters.  By default, any non-whitespace characters in
that group will also be deleted while aligning the
alignment character.  However, if the `justify' attribute
is set to a non-nil value, only the initial whitespace
characters within that group will be deleted.  This has
the effect of right-justifying the characters that remain,
and can be used for outdenting or just plain old right-
justification.

Alternatively the various table-editing options can also deal with this (e.g. org, ses, table-capture/release), or you could do it with an elisp replacement pattern.

e.g. The following should do more or less what you're looking for, provided that the file is already using spaces for alignment (you can use untabify to remove the tabs if not), and that all lines are the same length (i.e. trailing spaces are needed on some lines if the final column is of varying length).

C-M-% \([0-9]+\)\([[:space:]]+\) RET \,(format (concat "%" (number-to-string (1- (length \&))) "d ") (string-to-number \1)) RET