Hi there, This question is a development question relating to how I might solve or approach a particular regular expression problem using the Jakarta ORO package. Here we go... Given an arbitrary string Foo and a set of regular expressions or patterns Bar, I'm looking to for the most optimal way to find the subset of patterns in Bar which Foo matches. Optionally, I'd also like to be able to determine the "best fit" pattern in the subset, if it has more than 1 item. The trivial brute force approach obviously involves linearly checking each pattern in the set with performance of O(n). But I'm anticipating having to deal with a relatively large set of patterns so this doesn't work. Are there any features within Jakarta ORO which would allow me to implement a more optimal solution? Can anyone suggest how the components in ORO might be combined/assembled to acheive something like this? One other thing to bear in mind, is that there will likely be a great deal overlap between the patterns in the set. In essence many of the patterns will be more and more specific "child" variants of a given general "parent" pattern. Therefore anything matching a child would necessarily also match the parent. Can ORO be used to recognize these "inheritance" (I'm sure there's a better word!) relationships between patterns? If so, I'm thinking I might be able to construct some type of nested tree-like data-structure whereby checking the "parents" will allow me to quickly eliminate most of the set. Anyway I'm babbling now. Any responses, help or suggestions you can provide at greatly appreciated. Thanks in advance. Regards, Sasha Haghani Brightspark http://www.brightspark.com/ Toronto, Canada E: [EMAIL PROTECTED] T: 416.488.1999 x 241 F: 416.488.1988
