Regexp help

[minestrone]

That is how you lose 2 hours doing the most trivial of things that should be easy.

Cheers.

[NickFitz]

The spaces comeout as 0xa0 if the come from a 'real' space and 0x20 if the come from a  

I put the XML through JTidy to get the DOM.

Never even know there were 2 types of space and the debugger clearly never knew the difference.

Recon i'm jusy going to replace all the 0x20 to 0xa0s

Cheers.

Yup, 0xa0 (aka 160) is the non-breaking space. Therefore,

[\s\xa0]-[\s\xa0]

should do what you want

[minestrone]

The spaces comeout as 0xa0 if the come from a 'real' space and 0x20 if the come from a  

I put the XML through JTidy to get the DOM.

Never even know there were 2 types of space and the debugger clearly never knew the difference.

Recon i'm jusy going to replace all the 0x20 to 0xa0s

Cheers.

[minestrone]

Hyphen-minus is 0x2d, en dash is 0x2013. Time for a hex dump

just going to do that.

[minestrone]

Ahh, the ones that work have " - " in their data but the ones that fail have  - .

I am reading the XML data into a DOM then going through it but it looks like the 2 are not working the same even though they come out as spaces for both when I look at them in the debugger.

[NickFitz]

I have discovered the ones that work do come from a different list and I can repeat the fail on everyone from the other list. No idea why that is though, the chars look the same.

Hyphen-minus is 0x2d, en dash is 0x2013. Time for a hex dump

[minestrone]

System.out.println( teamNames );
String[] splitNames = teamNames.split("\\s-\\s") ;
System.out.println( splitNames.length );
System.out.println( splitNames[0] );

Gives

name1 - name2
1
name1 - name2

Arrghh

Even went as far as copying the dash out of the debugger and also from the console to make sure the thing was the same as I was looking for.

I have discovered the ones that work do come from a different list and I can repeat the fail on everyone from the other list. No idea why that is though, the chars look the same.

[NickFitz]

JDK 1.7.

It seems to work OK for 100 examples then just barfs on one through no other apparent reason that the regexp.

Ill try the others after me lunch.

Check that the failing one is really what it seems to be. If it has, say, an en-dash (–) instead of a hyphen (-)...

[minestrone]

JDK 1.7.

It seems to work OK for 100 examples then just barfs on one through no other apparent reason that the regexp.

Ill try the others after me lunch.

[NickFitz]

Which regex implementation are you using? \s-\s works for me™ in JDK, Jakarta-ORO Perl 5, Jakarta-ORO AWK, JRegex and Jakarta-Regexp implementations.

[Spacecadet]

try

[ ]-[ ]

Thats a space inside the square brackets


you can also use

[:space:]-[:space:]

I've not had any problems with the former method

[minestrone]

I'm trying to spit a string and it wants a regexp to specify the delimeter.

The delimeter should be " - " as in <space>-<space>.

I have tried \\s-\\s but that still gives me the odd error.

Anyone know how to exactly specify the delimeter?

Cheers.