On Thu, 12 Jun 2014 05:44:37 -0700, Amit Kumar Dev <[email protected]> wrote:
> Hi, > > I am using one character search with term "D*" in "class-number" node of > the xml. > I am getting docs which has class-number starts with "D" > In response I am also getting few documents which contains "D" after "/" > (slash) or "." (dot) for class-number element. > > Response for search Term "D*" in class-number element: > <class-number>D245/123<class-number>------------Expected > <class-number>A245/123 D5<class-number>--------Expected since space > separates the terms > <class-number>1245/D123<class-number>-----------Not Expected > <class-number>1245.D123<class-number>-----------Not Expected Actually, these are expected as well. Punctuation is a token break. > > I tried with below queries: > > * Element query with combination of word query. > * Element word query. > * Field word query by creating field for class-number element. All these are word queries, and after punctuation you are starting a new word. I see two options for you: (1) Recast your data so that you can use exact value queries. That is, your second case above would not be a match on a value query, so if that is a problem, you'd need to use multiple class-number elements or something. (2) (ML7) Use customized tokenization on a field over class-number to make the punctuation you expect to be part of the token into word characters so that you don't get the token break. //Mary _______________________________________________ General mailing list [email protected] http://developer.marklogic.com/mailman/listinfo/general
