You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@manifoldcf.apache.org by Frank Brendel <fr...@eurolog.com> on 2015/03/05 09:27:15 UTC

Move metadata changes content

Hello,

I am trying to index SharePoint documents in an OpenSearchServer index.
Therefor I have to move the meta data 'uri' coming from SharePoint to 
'url' required by the index schema.

But the meta data is not moved. It is copied into a new meta data and 
the content is changed.
As result I get two meta data 'uri' and 'url' with different content.

Here is what I have sniffed:

<field name="uri">
    <value>
http://share.point.server:1234/Site/SubSite/Documents/Document.docx
    </value>
</field>
<field name="url">
    <value>
       <![CDATA[/Site/SubSite/Documents/Document.docx]]>
    </value>
</field>

The problem is that the links presented by the search web page doesn't 
work because the "protocol://host:port" part is removed.
How can I keep the link intact?


Kind regards
Frank

Re: Move metadata changes content

Posted by Frank Brendel <fr...@eurolog.com>.
Hi Karl,

Am 05.03.2015 um 10:34 schrieb Karl Wright:
> The OpenSearchServer connector reserves "uri" to be the document's 
> actual URI, which in ManifoldCF's output connectors means the 
> document's key.  So you cannot override that.  Nor does it actually 
> come from a metadata field called "uri".  So, if I assume you are 
> trying to move that around using the Metadata Adjuster transformer, 
> you cannot actually describe that key and specify that it gets moved.
that's good to know.
And now it's clear why I see only a rewrite of the 'Attribute name' from 
the 'Path metadata' when I provide one and not of the 'uri'.

>
> But you *can* count on the uri field in OpenSearchServer, or the "id" 
> field in Solr, etc. to contain a usable URL that you can click on.  
> That's usually the way people do things.
First I started to bring MantisBT and a Windows Share together and 
defined the 'url' field as the key in the OpenSearchServer index schema. 
Ok, 'uri' is much better.
But now I am standing here like my friend Bert and have to integrate the 
SharePoint with all its high confidential documents ;-)

I think I start over and use Solr.


Thank you
Frank

>
> The "url" field that you describe *is* actually part of the metadata, 
> and comes directly from SharePoint.  You can adjust it in whatever 
> manner you see fit using the Metadata Adjuster transformer, including 
> adding on a prefix that would make it be a complete URL, if you choose.
>
> Hope this helps.
>
> Karl
>
>
> On Thu, Mar 5, 2015 at 3:27 AM, Frank Brendel 
> <frank.brendel@eurolog.com <ma...@eurolog.com>> wrote:
>
>     Hello,
>
>     I am trying to index SharePoint documents in an OpenSearchServer
>     index.
>     Therefor I have to move the meta data 'uri' coming from SharePoint
>     to 'url' required by the index schema.
>
>     But the meta data is not moved. It is copied into a new meta data
>     and the content is changed.
>     As result I get two meta data 'uri' and 'url' with different content.
>
>     Here is what I have sniffed:
>
>     <field name="uri">
>        <value>
>     http://share.point.server:1234/Site/SubSite/Documents/Document.docx
>        </value>
>     </field>
>     <field name="url">
>        <value>
>           <![CDATA[/Site/SubSite/Documents/Document.docx]]>
>        </value>
>     </field>
>
>     The problem is that the links presented by the search web page
>     doesn't work because the "protocol://host:port" part is removed.
>     How can I keep the link intact?
>
>
>     Kind regards
>     Frank
>
>

Re: Move metadata changes content

Posted by Karl Wright <da...@gmail.com>.
Hi Frank,

The OpenSearchServer connector reserves "uri" to be the document's actual
URI, which in ManifoldCF's output connectors means the document's key.  So
you cannot override that.  Nor does it actually come from a metadata field
called "uri".  So, if I assume you are trying to move that around using the
Metadata Adjuster transformer, you cannot actually describe that key and
specify that it gets moved.

But you *can* count on the uri field in OpenSearchServer, or the "id" field
in Solr, etc. to contain a usable URL that you can click on.  That's
usually the way people do things.

The "url" field that you describe *is* actually part of the metadata, and
comes directly from SharePoint.  You can adjust it in whatever manner you
see fit using the Metadata Adjuster transformer, including adding on a
prefix that would make it be a complete URL, if you choose.

Hope this helps.

Karl


On Thu, Mar 5, 2015 at 3:27 AM, Frank Brendel <fr...@eurolog.com>
wrote:

> Hello,
>
> I am trying to index SharePoint documents in an OpenSearchServer index.
> Therefor I have to move the meta data 'uri' coming from SharePoint to
> 'url' required by the index schema.
>
> But the meta data is not moved. It is copied into a new meta data and the
> content is changed.
> As result I get two meta data 'uri' and 'url' with different content.
>
> Here is what I have sniffed:
>
> <field name="uri">
>    <value>
> http://share.point.server:1234/Site/SubSite/Documents/Document.docx
>    </value>
> </field>
> <field name="url">
>    <value>
>       <![CDATA[/Site/SubSite/Documents/Document.docx]]>
>    </value>
> </field>
>
> The problem is that the links presented by the search web page doesn't
> work because the "protocol://host:port" part is removed.
> How can I keep the link intact?
>
>
> Kind regards
> Frank
>