Re: regex match only outside the double quotes

You have three cases:

Line start with --, you want to remove the entire line aka replace with nothing.
A substring starting and ending with " should be left alone, aka replace with itself.
-- elsewhere should have the part after the -- removed, aka, replace the entire thing with --.

This is trivial in a regexp:

s/("[^"]*")|^--.*\n|(--).*/$1||$2||""/egm;
[download]

Comment on Re: regex match only outside the double quotes Download Code

Replies are listed 'Best First'.
Re^2: regex match only outside the double quotes by SuicideJunkie (Vicar) on Aug 21, 2009 at 13:24 UTC
As long as you don't have escaped quotes... Something like `"App1 said \"Died -- boom!\" at 3:14am"` Would go badly. -SJ How to make perldoc.perl.org resizable too!	[reply] [d/l]
Re^3: regex match only outside the double quotes by JavaFan (Canon) on Aug 21, 2009 at 13:27 UTC
The OP never said anything about escaped quotes, which I why I didn't include them I can extend the regexp to include escapes in my sleep, which both hands tied behind my back. And I'd include loop unrolling.	[reply]
Re^4: regex match only outside the double quotes by SuicideJunkie (Vicar) on Aug 21, 2009 at 18:55 UTC
In many cases the simple way is just fine. In fact I've done the same sort of thing just today; splitting a CSV file on commas because I happen to know that it is safe for my particular input. But it is always good to be aware of the potential pitfalls so that you're not completely surprised when the bottom drops out because the input has changed or somebody typed a bad description.	[reply]
Re^5: regex match only outside the double quotes by JavaFan (Canon) on Aug 21, 2009 at 22:58 UTC