Question

I have a set of coordinates exported from google sketchup with extra fluff that I’ve been trying to strip with regex. I think it’s really interesting for quickly getting drawings in 3D from e.g. SketchUp into canvas from and .xsi file. The are multiple instances of data sets in one variable:

$str = 'SI_NurbsCurve Edge1 {
        1,
        0,
        0,
        4,
        0,0,1,1,
        2,
        870.243,1229.35,143.395,1
        927.537,1323.53,103.842,1
        }

        SI_NurbsCurve Edge2 {
        1,
        0,
        0,
        4,
        0,0,1,1,
        2,
        899.54,1217.88,116.255,1
        870.243,1229.35,143.395,1
        }';

I’ve attempted to remove everything from the multiple instances except the coordinate data with this regex:

$reg = '#SI_NurbsCurve Edge[^"]* {
        1,
        0,
        0,
        4,
        0,0,1,1,
        2,#';  
$rep=""; 
$str=preg_replace($reg,$rep,$str);

However, this result in only echoing the last coordinate set found in the string, in this example the following remains:

899.54,1217.88,116.255,1
870.243,1229.35,143.395,1

Besides that I’m trying to strip the last number “1” that occurs on each line of coordinates, so this entire example would end up looking like this:

870.243,1229.35,143.395,
927.537,1323.53,103.842,

899.54,1217.88,116.255,
870.243,1229.35,143.395,

I would be very grateful for your time and know-how!

Answer 1

Your first problem (getting only the last values) is probably caused by this:

#SI_NurbsCurve Edge[^"]*

You would need a non-greedy regex or if the value after Edge are just numbers:

#SI_NurbsCurve Edge[0-9]*

After that, you can chop of the last two characters of every remaining line.

You probably need to escape the { character as well: \{ and account for the } and spaces / new-lines after every set so the first line should be something like:

$str = '#(\}\s+)?SI_NurbsCurve Edge[0-9]* \{

See the working example (except for the last 2 chars of every line…) on Codepad.

To also get rid of the remaining ,1 at the end of each line, you can change the preg_replace line with:

$str=preg_replace(array($reg, '#,1\r#'),array($rep,"\r"),$str);

This works on Codepad at least but probably depends on the encoding of the newlines.