Hi Mister Florent,

Thks for your answer.
There is no symptom at all: no exception, no specific log, no deletion.
When I run the 'real' script on QConsole (or Oxygen connected through xdbc
to ML), it returns the focus after a while (whitout SVC-EXTIME) without any
deletion.
I'm trying with a colleague to simplify the script and recreate a readable
case for everyone.
For now it seems that the sem:sparql combined with sem:store has some
influences on this behavior

So I will come back soon with a simplified script.
Lanz


On Wed, Jan 31, 2018 at 6:26 PM, Florent Georges <li...@fgeorges.org> wrote:
>
> Salut Lancelot,
>
> I might have missed it, but what is the exact symptom? Any error message?
Have you tried the "real" script through QConsole? Any chance you simplify
it and post it here? More pairs of eyes might catch something you missed...
>
> Regards,
>
> --
> Florent Georges
> H2O Consulting
> http://h2o.consulting/
>
> On 30 Jan 2018 22:29, "Lanz" <lancelot.meuril...@gmail.com> wrote:
>>
>> Hi all,
>>
>> I've got this strange issue with xdmp:node-delete with a node in the
database with Marklogic 8.0-6.3.
>> First when I use xdmp:node-delete on the target node with the above
basic script, it works:
>> ------------------------------------------------------------------------
>> xquery version "1.0-ml";
>> declare namespace html = "http://www.w3.org/1999/xhtml";;
>> declare namespace l="http://www.oecd.org/ns/lambda/schema/";;
>> import module namespace db = "http://www.oecd.org/ns/lambda/app/lib/db";
at "/app/lib/db/db.xqm";
>> import module namespace exportlib = "http://www.oecd.org/ns/
lambda/app/lib/export" at "/app/lib/export/export.xqm";
>> declare variable $keyName-parent-toc-info as xs:string :=
'parent-toc-info';
>> let $rid := 'urn:oecd.org:publications:id:expression:g2g12781'
>> let $doc := fn:doc(db:get-uri-by-rid($rid))/l:expression
>> return
>> (
>> (: display parent node before child node deletion :)
>> $doc/l:exportChannels/l:exportChannel[@href =
$exportlib:channel-rid-i2])/l:precomputedValues
>> ,xdmp:node-delete($doc/l:exportChannels/l:exportChannel[@href =
$exportlib:channel-rid-i2])/l:precomputedValues/l:precomputedValue[@key =
$keyName-parent-toc-info]
>> )
>> ------------------------------------------------------------------------
>> The node 'l:precomputedValues[l:precomputedValue[@key =
$keyName-parent-toc-info]]' is removed from document in database
>>
>> But if I try to do it in the following script (the real script), it does
not work:
>> 1- I populate a map $mapA with ids of eligible documents based on a
sem:sparql query using a sem:store constructor
>> 2- I populate a map $mapB with ids of larger set of documents based on a
sem:sparql query using a sem:store constructor with a
cts:element-attribute-value-query on the element to be deleted
(cts:element-attribute-value-query
(xs:QName('l:precomputedValue'),xs:QName('key'),
$keyName-parent-toc-info)) among other cts:query constructor
>> 3- Then I create a 3rd map based on the difference of the two maps
($mapB - $mapA)
>> 4- I loop over this map of ids to get the uri of the document, then the
document with fn:doc(), then I get the element l:precomputedValues to be
deleted with an Xpath (the same as in the basic script)
>> 5- It never works on Oxygen using an xdbc connection on ML db or event
in Qconsole. There is no error, it takes a while and I've got nothing in
the error log (even with the finest setting). When I launch the profiler on
Qconsole, here the message I get at the end :
>> [1.0-ml] XDMP-CAST: (err:FORG0001) xs:unsignedLong($d div
xs:dayTimeDuration("PT0.000001S")) -- Invalid cast: -178609914.9 cast as
xs:unsignedLong
>>
>> I confirm:
>> - the node to be deleted is in the db not an on-the-fly constructed node
>> - the id used in the first successful script comes from the final map in
second script. This final map is populated with correct and real document
ids.
>>
>> Any help is appreciated
>> Lanz
>>
>> _______________________________________________
>> General mailing list
>> General@developer.marklogic.com
>> Manage your subscription at:
>> http://developer.marklogic.com/mailman/listinfo/general
>>
>
> _______________________________________________
> General mailing list
> General@developer.marklogic.com
> Manage your subscription at:
> http://developer.marklogic.com/mailman/listinfo/general
>
_______________________________________________
General mailing list
General@developer.marklogic.com
Manage your subscription at: 
http://developer.marklogic.com/mailman/listinfo/general

Reply via email to