Hi Rachel,
Just to give you an idea how you could replace a search with a lookup:
xquery version "1.0-ml";
import module namespace dls = "http://marklogic.com/xdmp/dls";
at "/MarkLogic/dls.xqy";
for $x in cts:search(xdmp:document-properties(),
dls:documents-query()
)
let $y := cts:element-values(xs:QName("dls:version-id"), (),
("properties"),
cts:and-query((
cts:element-value-query(
xs:QName("dls:latest"),"true",
(),0),
cts:element-value-query(
xs:QName("dls:document-uri"),
base-uri($x),
(),0)
))
)
return
<content>{$x/*:properties/*[not(namespace-uri() =
'http://marklogic.com/xdmp/dls')]}<version>{$y}</version></content>
Cheers,
Geert
> -----Oorspronkelijk bericht-----
> Van: Geert J. [mailto:geert.jos...@dayon.nl]
> Verzonden: vrijdag 7 februari 2014 0:59
> Aan: MarkLogic Developer Discussion
> Onderwerp: RE: [MarkLogic Dev General] Efficient way to get the version
> number of latest dls managed document
>
> Hi Rachel,
>
> Just to be sure, I compared the dls library between ML 6 and ML 7, but
> apart from some small changes related to external users (LDAP related
> things, new in ML 7), the library was pretty much unchanged.
>
> I ran a small test case, to see more clearly what happens. So, in short,
> there are always two copies having latest 'true', the main doc, and the
> latest version doc. And only the latter has indeed the version-id. So,
> yes, you will need to pull properties from two fragments together to get
> the result you are looking for.
>
> There isn't a join like you might be used to in SQL, but it is still
very
> well possible to pull things together. You will just need to do multiple
> searches, or if you are only looking for values, do multiple lexicon
look
> ups. The more blunt approach would be to do something like:
>
> for $x in cts:search(xdmp:document-properties(),
> dls:documents-query()
> )
> let $y := cts:search(xdmp:document-properties(),
> cts:and-query((
> cts:element-value-query(
> xs:QName("dls:latest"),"true",
> (),0),
> cts:element-value-query(
> xs:QName("dls:document-uri"),
> base-uri($x),
> (),0)
> ))
> )
> return
> <content>{$x, $y}</content>
>
> I guess it essentially resembles your approach. I can imagine such
> approach could get slow as the database grows.
>
> But as said, you could replace the inner cts:search by lexicon lookups
if
> you are only interested in the version-id, which is indexed afaik. And
if
> your other properties are atomic, you might be able to compose those
using
> lexicon lookups as well, by adding range indexes on them. That would
avoid
> disk io all together...
>
> Kind regards,
> Geert
>
> > -----Oorspronkelijk bericht-----
> > Van: general-boun...@developer.marklogic.com [mailto:general-
> > boun...@developer.marklogic.com] Namens Rachel Wilson
> > Verzonden: donderdag 6 februari 2014 14:01
> > Aan: MarkLogic Developer Discussion
> > Onderwerp: Re: [MarkLogic Dev General] Efficient way to get the
version
> > number of latest dls managed document
> >
> > Thanks for the reply Geert :)
> >
> > We're using ML6.
> >
> > The version-id property does exist but only on the "version" versions
of
> > documents (for want of a better way to put it). The "master" document
> > contains a property that we're interested in and our ultimate query
> result
> > needs to include the property on that master document as well as the
> > latest version number.
> >
> > <result>
> >
> > <uri>/project/lappingpublish/content/243e2f5f-f2c9-4f9d-89bb-
> > 491163cd0e00</
> > uri> (:this one is what we're calling the "master" document:)
> > <dls:version xmlns:dls="http://marklogic.com/xdmp/dls";>
> > <dls:latest>true</dls:latest>
> > </dls:version>
> > <prop:userDefinedProperty>interestingValue<prop:userDefinedProperty>
> > </result>
> > <result>
> >
> > <uri>/project/lappingpublish/content/243e2f5f-f2c9-4f9d-89bb-
> > 491163cd0e00_2
> >
43e2f5f-f2c9-4f9d-89bb-491163cd0e00_versions/1-243e2f5f-f2c9-4f9d-89bb-
> > 4911
> > 63cd0e00</uri>
> > <dls:version xmlns:dls="http://marklogic.com/xdmp/dls";>
> > <dls:version-id>1</dls:version-id>
> > <dls:latest>false</dls:latest>
> > </dls:version>
> > </result>
> > ...
> > <result>
> >
> > <uri>/project/lappingpublish/content/243e2f5f-f2c9-4f9d-89bb-
> > 491163cd0e00_2
> >
43e2f5f-f2c9-4f9d-89bb-491163cd0e00_versions/4-243e2f5f-f2c9-4f9d-89bb-
> > 4911
> > 63cd0e00</uri>
> > <dls:version xmlns:dls="http://marklogic.com/xdmp/dls";>
> > <dls:version-id>4</dls:version-id>
> > <dls:latest>true</dls:latest>
> > </dls:version>
> > </result>
> >
> >
> >
> > So we have a function at the moment that looks like below, but it's
> > getting slow for large projects because we need to run something
similar
> > to you els-latest-version function and map lookup hundreds of
thousands
> of
> > times. I guess I was hoping there was an optimised dls function
similar
> > to dls:documents-query() that I'd missed somewhere.
> >
> > Eg.
> >
> > declare function local:get-file-ids($projectId) {
> > let $baseUri := concat("/project/",$projectId,"/content/")
> > let $fileIdLookup := map:map()
> >
> > (: ML doesnt do joins between docs and so in order to lookup the
> fileid
> > property from master doc later we'll need a map. :)
> > let $x := for $docProp in
> cts:search(xdmp:directory-properties($baseUri),
> >
> > cts:properties-query(cts:element-value-query(xs:QName("prop:fileId"),
> "*",
> > "wildcarded")),
> > "unfiltered")
> > return map:put($fileIdLookup, base-uri($docProp),
> > $docProp/prop:properties/prop:fileId/text())
> >
> > (: find latest undeleted document including version number, can't
use
> > dls:documents-query() because version id only lives on versioned doc
:)
> > let $results := cts:search(xdmp:directory-properties($baseUri,
> > "infinity"),
> > cts:and-query(
> >
> > (cts:properties-query(cts:element-value-query(xs:QName("dls:latest"),
> > "true")),
> >
> >
cts:properties-query(cts:element-value-query(xs:QName("dls:version-id"),
> > "*", "wildcarded")),
> >
> > cts:not-query(cts:properties-query(cts:element-value-
> > query(xs:QName("prop:d
> > eletedBy"), "*", "wildcarded")))
> > )), "unfiltered")
> >
> > (: now join it all together :)
> > for $docProp in $results
> > let $mainDoc :=
> $docProp/prop:properties/dls:version/dls:document-uri
> > return <content>
> >
<id>{local:get-content-id-from-resultset($mainDoc/string())}</id>
> > <fileId>{map:get($fileIdLookup, $mainDoc)}</fileId>
> >
> >
>
<version>{$docProp/prop:properties/dls:version/dls:version-id/string()}</v
> e
> > rsion>
> > </content>
> > };
> >
> >
> >
> >
> >
> >
> > -----Original Message-----
> > From: "Geert J." <geert.jos...@dayon.nl>
> > Reply-To: MarkLogic Developer Discussion
> > <general@developer.marklogic.com>
> > Date: Thursday, 6 February 2014 07:41
> > To: MarkLogic Developer Discussion <general@developer.marklogic.com>
> > Subject: Re: [MarkLogic Dev General] Efficient way to get the version
> > number of latest dls managed document
> >
> > Hi Rachel,
> >
> > Apologies for the late reply. Looking at the dls library code suggests
> > that the version-id property should normally be present:
> >
> > declare private function dls-latest-version(
> > $uri as xs:string
> > ) as xs:unsignedInt
> > {
> > if (fn:exists(xdmp:directory-properties(version-directory($uri))))
> > then
> > let $latest :=
> > cts:search(
> > xdmp:document-properties(),
> > cts:and-query((
> > cts:element-value-query(
> > xs:QName("dls:latest"),"true",
> > (),0),
> > cts:element-value-query(
> > xs:QName("dls:document-uri"),$uri,
> > ("exact"),0))),
> > ("score-simple"),0)
> > return
> > if ($latest)
> > then $latest/prop:properties/dls:version/dls:version-id
> > else 0
> > else 0
> > };
> >
> > What version of MarkLogic are you running?
> >
> > Kind regards,
> > Geert
> >
> > > -----Oorspronkelijk bericht-----
> > > Van: general-boun...@developer.marklogic.com [mailto:general-
> > > boun...@developer.marklogic.com] Namens Rachel Wilson
> > > Verzonden: vrijdag 24 januari 2014 16:14
> > > Aan: MarkLogic Developer Discussion
> > > Onderwerp: Re: [MarkLogic Dev General] Efficient way to get the
> version
> > > number of latest dls managed document
> > >
> > > I was just wondering if anyone had any pointers in relation to this
> > > question. It's been a busy week on the list and it's only a small
> > > question I realise, but just in case it slipped through the net...
> > >
> > > -----Original Message-----
> > > From: Rachel Wilson <rachel.wil...@bbc.co.uk>
> > > Date: Wednesday, 22 January 2014 15:04
> > > To: MarkLogic Developer Discussion <general@developer.marklogic.com>
> > > Subject: Efficient way to get the version number of latest dls
managed
> > > document
> > >
> > > What is the most efficient way of getting back the version number of
> the
> > > latest dls managed document, in a cts:search or similar, given that
it
> > is
> > > not possible to join two documents?
> > >
> > > It would be easy if the current document actually showed the
> > > <dls:version-id>2</dls:version-id> element (just like the previous
> > > versions), but it doesn't. It only shows
<dls:latest>true</dls:latest>
> > >
> > > E.g. Dls properties of latest revision (what we tend to call the
> > "master"
> > > document) at /project/myproject/content/a
> > >
> >
> <http://bgb01xuev1003.national.core.bbc.co.uk/EnterpriseVault/ViewMessag
> > > e.a
> > >
> >
> sp?VaultId=18D759FC5FAC19A4B9ED689AE23969ADD1110000BGB01XUEV100
> > > 1&SavesetId=
> > >
> >
> 201312147013497~201309131531480000~Z~607DFE45035DC771F5921AC26A
> > > EFFFB1>:
> > > <dls:version xmlns:dls="http://marklogic.com/xdmp/dls";>
> > > <dls:latest>true</dls:latest>
> > > <dls:created>2013-09-13T15:13:57.225869+01:00</dls:created>
> > > <dls:author>10428144080895388018</dls:author>
> > > <dls:annotation/>
> > > <dls:deleted>false</dls:deleted>
> > > </dls:version>
> > >
> > >
> > > Dls properties of a previous version at
> > > /project/myproject/content/a_a_versions/1-a
> > >
> >
> <http://bgb01xuev1003.national.core.bbc.co.uk/EnterpriseVault/ViewMessag
> > > e.a
> > >
> >
> sp?VaultId=18D759FC5FAC19A4B9ED689AE23969ADD1110000BGB01XUEV100
> > > 1&SavesetId=
> > >
> >
> 201312147013497~201309131531480000~Z~607DFE45035DC771F5921AC26A
> > > EFFFB1>
> > > <dls:version xmlns:dls="http://marklogic.com/xdmp/dls";>
> > > <dls:version-id>1</dls:version-id>
> > >
> > > <dls:document-uri>/project/testapi-editing-
> > > activity/content/a</dls:document
> > > -uri>
> > > <dls:latest>false</dls:latest>
> > > <dls:created>2013-09-13T15:13:55.905791+01:00</dls:created>
> > > <dls:replaced>2013-09-13T15:13:56.787683+01:00</dls:replaced>
> > > <dls:author>10428144080895388018</dls:author>
> > > <dls:annotation/>
> > > <dls:deleted>false</dls:deleted>
> > > </dls:version>
> > >
> > >
> > > Specifically, I'd like to write a query to retrieve some user
defined
> > > property data from the properties of the "master" document as well
as
> > the
> > > version number of the latest version.
> > >
> > > This won't work because the properties for the document returned by
> > > dls:documents-query() doesn't have the version number.
> > > ---
> > > for $x in cts:search(xdmp:document-properties(),
> > > cts:and-query((
> > > dls:documents-query(),
> > > ))
> > > )
> > >
> > > return
> > > <content>
> > >
> > >
> >
>
<fileId>{$x:/prop:properties/prop:myuserdefinedproperty/string()}</fileId>
> > >
> > >
> >
>
<version>{$x/prop:properties/dls:version/dls:version-id/string()}</version
> > >
> > > </content>
> > > ---
> > >
> > > And this won't work because I also need the property in the "master"
> dls
> > > document that doesn't exist on the dls versions properties
> > > ---
> > > for $x in cts:search(xdmp:document-properties(),
> > > cts:and-query((
> > >
> > cts:element-value-query(xs:QName("dls:latest"),
> > > "true"),
> > >
> > > cts:element-value-query(xs:QName("dls:version-id"), "*",
> "wildcarded"),
> > > ))
> > > )
> > >
> > > return
> > > <content>
> > >
> > >
> >
>
<fileId>{$x:/prop:properties/prop:myuserdefinedproperty/string()}</fileId>
> > >
> > >
> >
>
<version>{$x/prop:properties/dls:version/dls:version-id/string()}</version
> > >
> > > </content>
> > > ---
> > >
> > > I suspect that some might suggest denormalising the user defined
> > property
> > > (would they?) but for the moment I'm stuck with the data structure
as
> it
> > > is.
> > >
> > > Many thanks,
> > > Rachel
> > >
> > >
> > >
> > >
> > > -----------------------------
> > > http://www.bbc.co.uk
> > > This e-mail (and any attachments) is confidential and
> > > may contain personal views which are not the views of the BBC unless
> > > specifically stated.
> > > If you have received it in
> > > error, please delete it from your system.
> > > Do not use, copy or disclose the
> > > information in any way nor act in reliance on it and notify the
sender
> > > immediately.
> > > Please note that the BBC monitors e-mails
> > > sent or received.
> > > Further communication will signify your consent to
> > > this.
> > > -----------------------------
> > > _______________________________________________
> > > General mailing list
> > > General@developer.marklogic.com
> > > http://developer.marklogic.com/mailman/listinfo/general
> > _______________________________________________
> > General mailing list
> > General@developer.marklogic.com
> > http://developer.marklogic.com/mailman/listinfo/general
> >
> >
> >
> > -----------------------------
> > http://www.bbc.co.uk
> > This e-mail (and any attachments) is confidential and
> > may contain personal views which are not the views of the BBC unless
> > specifically stated.
> > If you have received it in
> > error, please delete it from your system.
> > Do not use, copy or disclose the
> > information in any way nor act in reliance on it and notify the sender
> > immediately.
> > Please note that the BBC monitors e-mails
> > sent or received.
> > Further communication will signify your consent to
> > this.
> > -----------------------------
> > _______________________________________________
> > General mailing list
> > General@developer.marklogic.com
> > http://developer.marklogic.com/mailman/listinfo/general
_______________________________________________
General mailing list
General@developer.marklogic.com
http://developer.marklogic.com/mailman/listinfo/general
