XML is also a serialized form of the triples. Keep in mind MarkLogic has actual
data-types for triples, iris, and a lot more. Serializing them can be as simple
as document{ $triples }/*, or calling the appropriate sem function that
probably does something similar for you..
Kind regards,
Geert
From:
<[email protected]<mailto:[email protected]>>
on behalf of Florent Georges <[email protected]<mailto:[email protected]>>
Reply-To: MarkLogic Developer Discussion
<[email protected]<mailto:[email protected]>>
Date: Wednesday, November 18, 2015 at 8:44 PM
To: MarkLogic Developer Discussion
<[email protected]<mailto:[email protected]>>
Subject: Re: [MarkLogic Dev General] Managed vs Unmanaged Triples
Quick note: you don't need to 'serialize the triples to XML". You need to
express them as XDM (using specific element names). Which is anyway how managed
triples are stored, if I am right. Serializing is creating the textual
representation of that XDM, which should not happen if done right.
Yes, you can call me pedantic :-)
On 18 Nov 2015 19:21, "Geert Josten"
<[email protected]<mailto:[email protected]>> wrote:
Managed means they are stored/updated in MarkLogic automatically for you.
Unmanaged means you have to handle that yourself. If you just insert RDF data,
you end up with managed triples automatically. Same if you use SPARQL update.
So that is just more practical in that sense. In other words, it makes most
sense if you use MarkLogic as a plain triple store, or have no reason to embed
triples in documents.
To use documents and triples separately, just load the triples and the
documents separately, and maybe just have extra triples that have subject or
object value matching document uris to make links between the two worlds.
Regarding ’triples’, that wrapper has no real meaning for MarkLogic, but it can
be useful to bundle triples. It is used in managed triples documents as well,
and can also be useful to put annotations on. You could for instance add a
start-date and a end-date property on triples, and use that to get temporal
triples.
My best advice is to put triples in document properties. That means serializing
them to XML first, but that is just a minor detail for MarkLogic. Even if you
use DLS, you can still do direct updates on both documents, and properties. DLS
is just a library to help you with versioning, it is not enforced rigidly.
Though you probably want to provide some abstraction to make sure people do
their updates correctly..
Hope that helps!
Kind regards,
Geert
From:
<[email protected]<mailto:[email protected]>>
on behalf of "Robert Del Toro Jr."
<[email protected]<mailto:[email protected]>>
Reply-To: MarkLogic Developer Discussion
<[email protected]<mailto:[email protected]>>
Date: Wednesday, November 18, 2015 at 1:32 PM
To: MarkLogic Developer Discussion
<[email protected]<mailto:[email protected]>>
Subject: Re: [MarkLogic Dev General] Managed vs Unmanaged Triples
"If you are only interested in running SPARQL, nor want to annotate triples,
using managed triples makes most sense."
Why does it make the most sense? What is the advantage of doing it this way?
"If you are using document data as well, then you could still consider not
embedding them in the docs."
How do I include document data without embedding the triple in a document?
"A second option could be to store the triples in document properties, instead
of the body.
Consider using a triples property that has an array-node with triple objects."
I did notice before I went to bed last night that the example of embedded
triples in an xml doc on the MarkLogic documents was displayed inside a
'triples' object.
I guess I haven't really found a strong advantage to using managed triples,
which it seems like there should be (why do they call them managed anyway?),
but I'm still not sure that I like that I have to strip my triples out of my
document every time so the end user does not see them or that I have to update
the document just to update the triple (I wanted to use Document Library
Services to store changes to documents, but I did not want to store a new
version every time a triple was added, edited, or deleted so the embedded
triples seem like it will take more code to do that).
On Wed, Nov 18, 2015, 3:35 AM Geert Josten
<[email protected]<mailto:[email protected]>> wrote:
Hi Robert,
If you are only interested in running SPARQL, nor want to annotate triples,
using managed triples makes most sense. If you are using document data as well,
then you could still consider not embedding them in the docs. A second option
could be to store the triples in document properties, instead of the body.
I don’t think there are performance implications. The triple index will be
equally fast, and I have not heard that ingest speed of the triples will be
different because of this. I think it mostly comes down to what helps your
application best, which solution to select.
Regarding node-replace, that sounds like you are working with triple data
embedded in json. Consider using a triples property that has an array-node with
triple objects..
Kind regards,
Geert
From:
<[email protected]<mailto:[email protected]>>
on behalf of "Robert Del Toro Jr."
<[email protected]<mailto:[email protected]>>
Reply-To: MarkLogic Developer Discussion
<[email protected]<mailto:[email protected]>>
Date: Wednesday, November 18, 2015 at 7:21 AM
To: "[email protected]<mailto:[email protected]>"
<[email protected]<mailto:[email protected]>>
Subject: [MarkLogic Dev General] Managed vs Unmanaged Triples
Could someone please explain the pros and cons of using managed vs unmanaged
triples? I understand that with managed triples, I can update or delete the
triple without having to update the document, and with unmanaged triples I can
include extra information that I could then search with, however, when trying
to use triples stored in my documents I seem to run into complications. For
instance, when I try to read my document from an API, my triples come with it.
So in order to ensure that the user does not have access to the triples, I have
to use a transform with every request to remove the triples. Also, when trying
to update a document with more than one embedded triple using
'xdmp:node-replace', I get the error 'XDMP-CHILDDUPNAME: Object nodes cannot
have two children with the same name'. I'm assuming because I have multiple
nodes with the name 'triple'. Since using managed triples removes the ability
to add meta-data to the triples, what are the advantages? (Speed?)
Robert
_______________________________________________
General mailing list
[email protected]<mailto:[email protected]>
Manage your subscription at:
http://developer.marklogic.com/mailman/listinfo/general
_______________________________________________
General mailing list
[email protected]<mailto:[email protected]>
Manage your subscription at:
http://developer.marklogic.com/mailman/listinfo/general
_______________________________________________
General mailing list
[email protected]
Manage your subscription at:
http://developer.marklogic.com/mailman/listinfo/general
