Globally replace a particular pattern with another string

Question

In VI I need to replace a pattern that looks like this:

seq- followed by exactly 24 chars (can be 0-9, a-z, A-Z only) followed by / followed by some number of (0-9, a-z, A-Z). So an example string would be:

seq-BzFk78qrJBkbnv1KkxppsfG4/92191_A_B

I want to replace the above string with a fixed string:

seq-ID/NUM_ID

How do I do this in Vi or Vim? I tried a few things like

:%s/seq-[a-zA-Z0-9]{24}\/[a-zA-Z0-9]{+}/seq-ID\/NUM_ID

But it doesn't work.

Answer 1

There are many problems with your regex.

1. {24} will not match 24 of the previous atom, it will match the text {24}. I doubt this is what you want. You need a backslash e.g. \{24}. Additionally you could enable "Very magic", e.g. \v{24}

2. In your example, the text after the slash has underscores in it, e.g. 92191_A_B. Change [a-zA-Z-0-9] to [a-zA-Z-0-9_]

3. "One or more" is \+, not {+}.

There are also some style problems.

1. If you are going to search for something containing slashes, you should use a different separator, e.g.

   :%s;seq-[a-zA-Z0-9]{24}/[a-zA-Z0-9]{+};seq-ID/NUM_ID
   

   For more readability.

2. Don't use [a-zA-Z0-9_] when you could just use \w. I would also use this in place of [a-zA-Z0-9] unless it's possible that underscores will be in there, and you want to avoid matching this. In this case, \c[a-z0-9] is also shorter

Putting this all together, I would recommend going with:

:%s;\vseq-\w{24}/\w+;seq-ID/NUM_ID

I would also highly recommend reading through vim regex. One of the best quick intros to vim's regex engine I have ever seen. I have it bookmarked in my browser at all times for quick reference.