One suggestion is to run your schemas through a conversion process to eliminate the 1.1isms but keep as much of the typing . ML doesn't require documents to be schema valid unless you explicitly validate them. But it will use what it can from schemas to assign types to simple content.
For many things I've simply taken a corpus of a few docs (or few hundred) and ran them *with no schema* through a schema generation tool (like Oxygen has) and used the generated schema without even looking at it. It won't be perfect, and the more docs you give it the better - but even a few sample docs can usually deduce basic atomic types close enough... Alternatively you could use similar tools to down-convert 1.1 XSD to 1.0 and likely end up with a very accurate schema that's only missing a few things - like the asserts - which have no bearing at all on typing unless you use dynamic type assignment in the schema ... ( overstepping my knowledge here - I'm not sure there is such a thing). ----------------------------------------------------------------------------- David Lee Lead Engineer MarkLogic Corporation [email protected] Phone: +1 812-482-5224 Cell: +1 812-630-7622 www.marklogic.com<http://www.marklogic.com/> From: [email protected] [mailto:[email protected]] On Behalf Of Timothy W. Cook Sent: Friday, October 31, 2014 1:07 PM To: MarkLogic Developer Discussion Subject: [MarkLogic Dev General] datatypes and Range Indexes So, since my schemas are not valid because they use some XML Schema 1.1 features. I thought maybe I could get datatyping by building range indexes on certain elements. However, unless I am doing something wrong; range indexes do not help with this issue. I defined an Element Range index on an element, magnitude with the namespace http://www.mlhim.org/ccd and a type of xs:decimal. This query: declare namespace ccd = "http://www.mlhim.org/ccd";; for $doc in fn:doc() return xdmp:describe($doc//ccd:el-7ef40847-ea17-435e-856a-d489a32514d7/ccd:magnitude/data()) returns: xs:untypedAtomic("998070") So, does this mean that all of my queries will have to explicitly coerce the data? Which leaves me with fn:number() correct? now way to test that it is an integer, etc. Any guidance is appreciated. ============================================ Timothy Cook LinkedIn Profile:http://www.linkedin.com/in/timothywaynecook MLHIM http://www.mlhim.org<http://www.mlhim.org/>
_______________________________________________ General mailing list [email protected] http://developer.marklogic.com/mailman/listinfo/general
