I'm attempting to replace all strings in a text file surrounded by single quotes and ending with an @ first followed by any number and then any mix of any other letter(s) or #(s) like these:

'[email protected]'
'otherpackage@14'
'[email protected]'

and I wish to select only the single quotes and remove them like so:

export MYVAR="/dir/'[email protected]':$MYVAR" 

to:

export MYVAR="/dir/[email protected]:$MYVAR"

in the entire file. I figured out a way to remove the proceeding single quote and preceding single quote in two separate commands (using vim's zs and ze, similar to positive and negative lookarounds):

:%s/'.*@.[0-9]*.*\zs'//g 
:%s/\zs'\ze.*@.[0-9]*.*//g

However, I am curious about doing it in a single operation as I want to learn more about using Neovim, and apply the answer to future operations.

I am actually using Neovim, if that comes with any additional features related to this.

CodePudding user response：

You can replace matches of the following regular expression with the content of capture group 1.

'([^@']*@\d[^@']*)'

Demo

I'm not familiar with Vim's regex engine, but I understand it is quite robust. Considering that the regular expression I have suggested is nothing fancy (just making use of a capture group) I'm confident it should work in Vim.

The regular expression can be broken down as follows.

'           # match a single quote
(           # begin capture group 1
  [^@']*    # match zero or more characters other than '@' and a single-quote
  @\d       # match '@' followed by a digit ofollowed by zero or more characters
            # other than '@' and a single-quote
  [^@']*    # match zero or more characters other than '@' and a single-quote
)           # end capture group 1
'           # match a single quote

I've assumed that there can be at most one '@' between asterisks, but if more that one is permitted (if, for example, 'a@11@22*' is to be converted to a@11@22*) change the regular expression to

'([^']*@\d[^']*)'

CodePudding user response：

You need to use

%s/'\([^@'0-9]*@[0-9][^']*\)'/\1/g

Details:

• ' - a single quote
• \( - start of a capturing group
  • [^@'0-9]* - zero or more chars other than @, ' and digits
  • @ - a @ char
  • [0-9] - a digit
  • [^']* - zero or more chars other than a ' char
• \)' - end of the capturing group.

The replacement is \1, the Group 1 backreference/placeholder.