It’s a deceptively simple programming task. How do you capture a string of characters between single or double quotes?

Several times I’ve needed to grab name / value pairs of data entered by the user. For example, title="This is the Title", weight="28", etc. Using regular expressions, this would seem to be straightforward. The regular expression would simply match a string of characters in a manner similar to this:

(\w+)="([^"]*)"

Here’s a visual representation of what the above regular expression matches:

The Problem

However, what if our string of characters looks like this?

title="A Review of \"A Tale of Two Cities\""

Here the string of characters we want to capture itself has quotes in it; these quotes have been “escaped” using the backslash character \. What we want to “grab” out of the above string is everything between the first and last quotes, so we get the string A Review of "A Tale of Two Cities". However, our original regular expression won’t grab that. It only matches up to the first quote character it encounters: the string A Review of \.

Obviously this is problematic, because we need a regular expression that grabs as many “escaped characters” (that is, a backslash followed by another character) as possible. It needs to work even with something like this:

title="A Review of \"A Tale of Two Cities\" and \"Moby Dick\""

The Solution

The trick is to divide the string into consistently divisible parts. The string above, for example, divides into an (optional) initial string of characters, followed by repeated groupings that each start with \", like so:

By constructing a regular expression that matches this sequence zero or more times, we now have a solution. Our original regular expression, now modified, looks like this:

(\w+)="([^"\\]*(\\.[^"\\]*)*)"

Represented visually:

I’ve created a command-line program, written in PHP, that demonstrates this in action.