Text Parsing with Regex

[Andy Bowes]

Hi

I am trying to find a generic method to extract information from strings like:

"id:36 sub:001 dlvrd:001 submit date:0704270919 done date:0704270919 stat ELIVRD err:000"

I would like to be able to extract the values of particular items by defining a Regex style pattern. I am a bit of a regex newbie and I am not sure how to go about this. I had thought about using StringTokenizer but some of the value names have spaces in them and unfortunately I have no control over the format of these strings.

Can you use regex expressions to 'name' specific variable sections of the patterns so that they can be extracted by a Matcher?

Any help will be very welcome, thanks

[Omer Haderi]

Hi,

The java.util.Scanner has much more capabilities than StringTokenizer. Have a look at that class, you can delimited with regex also, moreover it has methods to return specific types.

Cheers.

[Andy Bowes]

Hi Omer

Thanks for the reply.
I have worked it out using Regex and it works like a dream

Andy

[Stan James]

How did you handle the names with spaces in them? I have to guess the values cannot have spaces, or you'd never know where a value ends and a name begins.

[Andy Bowes]

A Regular Expression can include spaces in the Pattern e.g.

String patternText = "^id :( [A-Za-z0-9]{1,10}) sub :( \\d{3}) dlvrd :( \\d{3}) submit date :( \\d{10}) done date :( \\d{10}) stat :( [A-Z]{1,8}) err :( \\d{3}) Text:.*"
 
   Pattern pattern = Pattern.compile( patternText );
   Matcher matcher = pattern.matcher(textToParse);
 
   //Extract the groups from the regex (e.g. elements in braces)
   String messageId = matcher.group( 1 );
   String submitDate = matcher.group(4);
   String statusText = matcher.group( 6 );
 

[ May 02, 2007: Message edited by: Andy Bowes ]

[Stan James]

Oh, cool, you made a pattern with all known keys and got them all with one match. I was thinking of something generic to get one key:value at a time which would survive insertion of a new key:value one day but wouldn't be near as self documenting.