Question

I need help to build a regular expression. I’m programming a software that looks for certain words in a string. But I need to also look for phrases. This string is entered by the user in a text box.

Currently I use the following regular expression \s+ to replace spaces by ampersands:
word1 word2 word3 to word1&word2&word3

My new requirement must include phrases enclosed in quotes, like the searching for phrases in Google.

"word1 word2" "word3 word4" word5

Must be:

word1 word2&word3 word4&word5

Thanks in advance.

EDIT: If there is another way or another approach to do the same, I’m open to any idea.

Answer 1

This assumes the quests are well balanced – you have an even number of quests, and not quests in middle of words.

You can match for /"([^"]+)"|(\S+)/ – this will result in your words – quoted strings or non-spaces. It captured the word in group 1 or 2, and you can then join the results with the & delimiter.

Another option: you can get it in a single replace, by which is very similar way to the first pattern, by skipping over your tokens. Here’s a JavaScript example:

s = s.replace(/(?:"([^"]+)"|(\S+))\s*/g, '$1$2&');

(note that you will have an extra ampersand in the end of the string)