I'm trying to create a regex to tokenize a string. An example string would be.
John Mary, "Name=blah;Name=blahAgain" "Hand=1,2"
I'm trying to get back:
- John
- Mary
- Name=blah;Name=blahAgain
- Hand=1,2
views:
129answers:
5
+1
A:
For that specific example, I would do:
([^\s]*)\s+([^,\s]*)\s*,\s*"([^"]*)"\s+"([^"]*)"
update: modified to split Mary and John
slebetman
2010-01-19 18:37:28
Thanks for the advice, but I'm using the Scanner class in java and it doesn't seem to like it.
binarymelon
2010-01-19 18:46:15
@slebetman: Mary and John aren't split with your regexp
Antony Hatchkins
2010-01-19 19:02:32
That's strange. I'm quite sure that the above is fairly plain old-school regexp, without any weird PCRE or egrep stuff. Are you sure you've escaped the `"` with `\"` in java?
slebetman
2010-01-19 19:03:00
@Anthony: In the original question Mary and John was specified as a single token: they were not meant to be split. Obviously the spec have changed. Sigh.. never expected feature creep on SO
slebetman
2010-01-19 19:04:38
I'm fairly sure I have it in correctly. Using the following string [Pattern regex = Pattern.compile("([^,]*)\\s*,\\s*\"([^\"]*)\"\\s+\"([^\"]*)\"");]
binarymelon
2010-01-19 19:14:14
Sorry, don't know java. Perhaps you should add java to your question tag to catch the eyes of java coders.
slebetman
2010-01-19 19:29:47
The problem is that it doesn't catch John. You need another group in the beginning, like so: ([^ ]+) ([^,\s]*)\s*,\s*"([^"]*)"\s+"([^"]*)"
ferdystschenko
2010-01-19 20:36:42
@ferdystschenko: Aha, good catch. Fixed.
slebetman
2010-01-19 23:29:45
This would get the comma after Mary as well as the quotes. Also it wouldn't catch all desired fields at once.
ferdystschenko
2010-01-19 20:05:09
Actually, it catches even only a single character each time, unless you put the '+' inside the parantheses.
ferdystschenko
2010-01-19 20:07:47
A:
Since you're using Java, why not use StringTokenizer? E.g.:
StringTokenizer st = new StringTokenizer("String to tokenize", " ");
while (st.hasMoreTokens())
{
// get next token
String someVariable = st.nextToken();
}
Chris
2010-01-19 20:01:25
A:
This works for your example:
(\w+) (\w+), \"([^"]+)" \"([^"]+)
Do all your string have exactly the same pattern?
ferdystschenko
2010-01-19 20:15:52
A:
One possible way: split at , followed by a space or at one of space or quotation mark:
"John Mary, \"Name=blah;Name=blahAgain\" \"Hand=1,2\"".split(",\\s|[\\s\"]")
Fabian Steeg
2010-01-19 23:53:07