views:

45

answers:

2

I've got the string

$result = "bei einer Temperatur, die etwa 20 bis 60°C unterhalb des Schmelzpunktes der kristallinen Modifikation"

which comes straight from a MySQL table. The table, and the php headers are both set to UTF-8

I want to strip the 'degree' symbol: http://en.wikipedia.org/wiki/Degree_symbol and replace it with the word 'degrees' to get:

"bei einer Temperatur, die etwa 20 bis 60degreesC unterhalb des Schmelzpunktes der kristallinen Modifikation"

but I can't get it to work with preg_replace.

If I do:

$result = preg_replace('/\xB0/u'," degrees ", $result ); - I get an empty string

And if I do::

$result = preg_replace('/\u00B0/u'," degrees ", $result ); - I get the error:

Warning: preg_replace() [function.preg-replace]: Compilation failed: PCRE does not support \L, \l, \N, \U, or \u at offset 1 in /var/www/html/includes/classes/redeyeTable.inc.php on line 75

I'm not great with encodings... what am I doing wrong here?

+1  A: 

If you use the 'u' modifier, the pattern is supposed to be treated as utf-8, so why not simply write '°' instead of '\u00B0' or '\xB0'?

greg0ire
$result = preg_replace('/°/u'," degrees ", $result ); does work.... why doesn't it work giving the string in hex?
Ed Lewis
+1  A: 

Use

$result = preg_replace('/\x{00B0}/u'," degrees ", $result );

Please see here for more information on the \x{FFFF}-syntax.

It's important to note the difference between \xB0 and \x{00B0}:

  • \xB0 denotes a single character with hex-code B0 (176 decimal) which is the degree symbol (°) in ISO-8859-1 for example
  • \x{00B0} denotes the unicode codepoint U+00B0 which describes the degree symbol (°) in the unicode system. This codepoint will be encoded using two bytes \xC2\xB0 when using UTF-8 encoding.
Stefan Gehrig
That works! Thank you Stefan and everyone who contributed.My mistake was not using the { } around the unicode codepoint. I appreciate the difference between \xB0 and \x{00B0} - it was more desperate trial and error that had me settling for \xB0 replacement on a unicode string.Stackoverflow once again is a life saver!
Ed Lewis
@Ed: You could mark the answer as "accepted" to show other users that this is the solution to your problem.
Stefan Gehrig