replace quot inside quot

[simon1663]

Hi there,
I am very poor at regular expression 🙁

Can't just get the hang of it...

Anyway, I have to replace quotes inside a quote ..

For example, consider the following string:

"I say, \"my name is \"simon1663\". hehe \""

If it was printed on screen it will look like :

I say, "my name is "simon1663". hehe"

I want to replace quotes inside the quote with \" . So, desired output is:

I say, "my name is \"simon1663\". hehe"

In another words want to replace all occurance of quotes inside a quote with \\\"

 Can it be done? Any help is greatly appreciated.

[DigitalExpl0it]

Did you try somthing like this

//change the value to somthing else

//$string = "This is a test "this is a test in quotes", this is a test!!! ";

echo "Unmodified string: $string<br>";

$string = ereg_replace( '"', '<-quote->', $string); 

echo "Modified string: $string<br>";

//now change the value back to what it was

$string = ereg_replace( '<-quote->', '"', $string); 

echo "Modified string back: $string";

// Strip the slashes

$string = stripslashes ($string);

echo "<br>$string";

[simon1663]

No no no!

I want to replace quotes inside the quotes with \\"

So, from the example I had earlier - the true string representation I want is like:

"I say, \"my name is \\"simon1663\\". hehe\""

[Weedpacket]

This is a bit confusing.

You have:

"I say, "my name is "simon1663". hehe""

And you want to turn it into

"I say, \"my name is \\"simon1663\\". hehe\""

Is that right? How are we (or, more importantly, how is the computer) to know that you don't want:

"I say, \"my name is \"simon1663\". hehe\""

which is also "quoting quotes inside quotes" - parsing it as:

"I say, ("my name is ")simon1663(". hehe")"

instead of

"I say, ("my name is ("simon1663"). hehe")"

However you cut it, a regexp won't do the job for you - there are some inherent limits on what they're capable of, and you've just run up against them. (Tell the truth, I can make out one possibility, but it depends critically on the answer to problem above. It probably won't be any simpler than the suggesion below, though.)

My suggestion would be to do a good old-fashioned scan through the string, outputting characters into another string as you read them, and keeping count of how deeply nested in quotes you are, so that you can output the requisite number of \ characters whenever you see a ".

Regexps can't count.

[Bunkermaster]

you might wanna have a look at AddCSlashes
http://www.php.net/manual/en/function.addcslashes.php

[simon1663]

The outmost quotes are just to indicate that it is a string. Since it has caused confusion - I will take it out now

So the actual string is:
I say, \"my name is \"simon1663\". hehe\"

I just want to replace the quotes around simon1663 with \\" so it looks like
I say, \"my name is \\"simon1663\\". hehe\"

The main application for this is to generate java codes from XSLT. How and what is it - that's a very long history. Please don't ask me to describe it - otherwise I will take hours to explain it:o

I am pretty sure it can done using regexp. E.g.

$str = "I say, \"my name is \"simon1663\". hehe\""

$result = ereg_replace( ".\".(\").\"." , "\\"" , $str);

I know this is not correct, but it will be something like this. E.g.

whatver - then - quote - then - whatever - then - QUOTE - then - whatever - then quote - whatever

And I want to replace the QUOTE inside the other two quotes with "\\""

Can it be done?
😕

[Weedpacket]

I still need a bit of clarification. Given a string, how does one know which quotes get extra slashes added? You say "inside quotes", but that doesn't help, since any given quote (except the two extreme ones) can be considered both as being inside quotes and marking the end of a pair of quotes. I just need to know how to recognise which case applies for any given \".

In other words, I still don't know the difference between

I say, (\"my name is \")simon1663(\", hehe\")
---slashed as:
I say, \"my name is \"simon1663\", hehe\"

and

I say, (\"my name is (\"simon1663\"), hehe\")
---slashed as
I say, \"my name is \\"simon1663\\", hehe\"

Let alone how you mean to parse:

My textbook says \"The Beatles performed \"You Say \"Goodbye\" and I say \"Hello\"\"\".

[simon1663]

You kind of turning it into more complex than what I am after.

Ok I will take a simpler approach.

Quetion is, write a regular expression for the use of ereg_replace which replace all occurance of quotes inside a quote with \\"

sample input #1 : I said, \"you said , \"hello\"\"
sample output #1 : I said, \"you said , \\"hello\\"\"

sample input #2 : I said, \"you said , \"hello\" but a quote \"?\"
sample output #2 : I said, \"you said , \\"hello\\" but a quote \\"?\"

Note that, all quotes inside the quote are replaced with \\". There is no matching quote concept here other than the end ones. So,

sample input #3: I said quotes, \"quote\"quote\"quote\"quote\"qoute\"done\"
sample output #3: I said quotes, \"quote\\"quote\\"quote\\"quote\\"qoute\\"done\"

Thanks for your help.

[Weedpacket]

Thanks, Now it makes sense.

I won't do this with ereg_replace. preg_replace offers a much more sophisticated syntax - you can do things with two characters in PCRE regexps that require exponentially long POSIX regexps.

Okay, so we don't mess with the the two end \" sequences, but we replace all the other \" sequences in between with \\".

If you could get the regexps right it could be done in three lines - the first to turn all \" into \\" and then two regexps to change the ones on the end back. I had a crack but kept losing count.

It could be done, but I still maintain it would be less eye-twisting to do it without regexps. (At least, I got this working):

$b = chr(92);
if(strpos($string,$b.'"')!==false) // Why bother, otherwise?
{	$string = str_replace($b.'"',"$b$b$b".'"',$string);
	$left = strpos($string, "$b$b$b".'"');
	$string = substr($string, 0, $left).substr($string,$left+2);
	$string = strrev($string);
	$right = strpos($string, '"'."$b$b$b");
	$string = substr($string, 0, $right+2).substr($string, $right+4);
	$string = strrev($string);
}

You'll note I went to some lengths to avoid having to use literal backslashes at all.

[Weedpacket]

No fuss, very little muss:

$b=chr(92);
$string = explode($b.'"',$string);
$count = count($string);
$string = $string[0]
	. $b . '"'
	. join("$b$b$b".'"', array_slice($string, 1, $count-2))
	. $b . '"'
	. $string[$count-1];

[simon1663]

Thanks!

I will try it....

Altough I would have really loved a regexp solution 😛

Anyway, thanks again and take care..

[Weedpacket]

Okay, a solution that uses regexps:

$string = 'This is \"a \"quoted\" string\". ';

$sq = chr(92).chr(34);
$sssq = chr(92).chr(92).chr(92).chr(34);

preg_match("/^((?:(?!$sq).)*)$sq(.*)$sq((?:(?!$sq).)*)\$/",$string,$match);
$string = $match[1].'"'.str_replace($sq,$sssq,$match[2]).'"'.$match[3];