If the dc:identifier is always either a URL or a Journal Title, you could probably use an xsl:choose statement, something like this:

 

                <!-- URL or Journal Title -->

               <xsl:for-each select="//dc:identifier">

                   <xsl:choose>

                       <xsl:when test="substring(., 1, 4) = &quot;http&quot;">

                           <field name="url">

                               <xsl:value-of select="." />

                           </field>

                       </xsl:when>

                       <xsl:otherwise>

                           <field name="container_title">

                               <xsl:value-of select="." />

                           </field>

                       </xsl:otherwise>

                   </xsl:choose>

               </xsl:for-each>

 

Notes:

 

1.)    That’s untested code, but hopefully it gives you the general idea.

2.)    I’m using the “container_title” field for storing journal titles in this example.  That field will be officially introduced in the forthcoming VuFind 1.4 release, but you can add it to an earlier version by looking at the articles2.patch file here: http://vufind.org/jira/browse/VUFIND-258

3.)    If you find a solution that works well, please consider adding some notes to the wiki: http://vufind.org/wiki/importing_records:how_to_index_eprints_with_vufind

 

I hope this is helpful; please let me know if you have any questions.

 

- Demian

 

From: Haik Zargaryan [mailto:haikzargaryan@yahoo.com]
Sent: Saturday, July 07, 2012 6:29 AM
To: vufind-general@lists.sourceforge.net
Subject: [VuFind-General] harvesting from e-prints

 

Hello dear community,

 

Our library has recently started exploring vufind. Everything works fine with one exception.

 

We have a huge digital content on e-prints and now we would like to import those records into our vufind catalog. 

This document http://vufind.org/wiki/importing_records appeared to be an excellent guide and we managed the task. Still we are not able to solve one task.

Our *e-prints record* contains two dc:identifiers; one for the URL and the other one for the name of the journal. Nonetheless, when we harvest the record and search it in the vufind, these two fields are not displayed (but they are very important for us). 

 

We understand, that this to be done we should make some modifications in the *dspace.xsl*, and maybe some other changes in the files responsible for the record display, but we do not know exactly what to do. 

 

Any advices?

 

Thanks in advance,

Haik.