Remove everything except match of regexp?

Question

I have a file with a structure like:

some text and numbers and a final number: 1.213,31
some other number 1231 and text 

again some text and a final number: 53,00
something else
again a final number: 1.233,54

Now I want to create a file looking like:

1.213,31
53,00
1.233,54

I can use higlight-regexp and the regexp [0-9,.]+$ to match the correct content, but I don't know how to remove the rest of the file or alternatively copy everything higlighted by this regexp into another buffer.

So how can I do this?

Answer 1

Just use query-replace-regexp with .*?\([0-9\,\.]+\)$ → \1. And then keep-lines with the same regexp.

Answer 2

alternatively copy everything [matching] this regexp into another buffer.

You can do that with occur

C-uM-so : \([0-9,.]+\)$ RET \1 RET

Or, if less-specific is fine, just:

C-uM-so [0-9,.]+$ RET