Hi,
we are using ML 7.0-2.3 and I have a query where sorting creates
XDMP-EXPNTREECACHEFULL. To tune the sorting we have already created a range
index on the sorting element
let $t := xdmp:query-trace(true())
let $res := (for $result in cts:search(/(standard|standard-enquiry),
cts:element-range-query(xs:QName("mle:xmltype"), "=", "std"))
order by $result//doc-number, $result//part-number
return $result)[1 to 10]
return $res
and the trace output shows no constraint is used for the sort
Analyzing path for search: fn:collection()/(standard|standard-enquiry)
Step 1 is searchable: fn:collection()
Step 2 is searchable: (standard|standard-enquiry)
Path is fully searchable.
Gathering constraints.
Step 2 contributed 1 constraint: standard
Step 2 contributed 1 constraint: standard-enquiry
Comparison contributed string range value constraint: mle:xmltype = "std"
Search query contributed 1 constraint:
cts:element-range-query(xs:QName("mle:xmltype"), "=", "std",
("collation=http://marklogic.com/collation/";), 1)
Executing search.
Selected 40694 fragments to filter.
but when I change the searchable expression to /standard I see that it uses the
range index for sorting
Analyzing path for search: fn:collection()/standard
Step 1 is searchable: fn:collection()
Step 2 is searchable: standard
Path is fully searchable.
Gathering constraints.
Step 2 contributed 1 constraint: standard
Comparison contributed string range value constraint: mle:xmltype = "std"
Search query contributed 1 constraint:
cts:element-range-query(xs:QName("mle:xmltype"), "=", "std",
("collation=http://marklogic.com/collation/";), 1)
Order by clause contributed 1 range ordering constraint for $result: order by
$result/descendant::doc-number ascending
Executing search.
Selected 40602 fragments to filter.
Is there a way to rephrase the searchable expression to make use of the index?
Will MarkLogic 8 handle this differently?
Thanks in advance for any input on that
Holger Apel
_______________________________________________
General mailing list
[email protected]
Manage your subscription at:
http://developer.marklogic.com/mailman/listinfo/general
