Just for reference, I figured how how to do the OCLC number reformatting in XSLT.

If we have a file 'format_oclc.marcxml':

<?xml version="1.0" encoding="UTF-8"?>

<record

xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"

xsi:schemaLocation="http://www.loc.gov/MARC21/slim http://www.loc.gov/standards/marcxml/schema/MARC21slim.xsd"

xmlns="http://www.loc.gov/MARC21/slim">

<subfield code="w">(OCoLC)2776588</subfield>

<subfield code="w">(OCoLC)12776588</subfield>

<subfield code="w">(OCoLC)112776588</subfield>

</datafield>

</record>

And an xslt file 'format_oclc.xslt':

<!DOCTYPE stylesheet >

<xsl:stylesheet version="1.0" xmlns:marc="http://www.loc.gov/MARC21/slim" xmlns:xsl="http://www.w3.org/1999/XSL/Transform" >

<xsl:template name="format_OCLC_number">

<xsl:param name="controlnumber"/>

<xsl:variable name="OCLC_number" select="substring-after($controlnumber,'(OCoLC)')"/>

<xsl:variable name="OCLC_length" select="string-length($OCLC_number)"/>

<xsl:if test="$OCLC_number">

<xsl:choose>

<xsl:when test="$OCLC_length < 8">

<xsl:variable name="OCLC_number_padding" select="substring( '00000000', 1, 8 - $OCLC_length)"/>

<xsl:variable name="formatted_OCLC_number" select="concat( $OCLC_number_padding, $OCLC_number )"/>

<xsl:value-of select="concat( 'ocm', $formatted_OCLC_number )" />

</xsl:when>

<xsl:when test="$OCLC_length = 8">

<xsl:value-of select="concat( 'ocn', $OCLC_number )" />

</xsl:when>

<xsl:otherwise>

<xsl:value-of select="concat( 'on', $OCLC_number )" />

</xsl:otherwise>

</xsl:choose>

</xsl:if>

</xsl:template>

<xsl:template match="marc:record">

<xsl:for-each select="marc:datafield[@tag=773]">

<xsl:for-each select="current()/marc:subfield[@code='w']">

<xsl:call-template name="format_OCLC_number">

<xsl:with-param name="controlnumber" select="current()"/>

</xsl:call-template>

<xsl:value-of select="'

'" />

</xsl:for-each>

</xsl:template>

</xsl:stylesheet>

we can run xsltproc:

$ xsltproc format_oclc.xslt format_oclc.marcxml

<?xml version="1.0"?>

ocm02776588 ocn12776588 on112776588

I'll be filing a bug/patch relatively soon that incorporates this.

On Sat, May 26, 2018 at 4:58 AM, Katrin Fischer <katrin.fischer.83@web.de> wrote:

I don't know about other sources, but in Germany I've never encountered data where the numbers in $w minus MarcOrgCode don't match the one in 001. Maybe this problem is just specific to OCLC?

Katrin

On 25.05.2018 20:06, Barton Chittenden wrote:

On Thu, May 24, 2018 at 3:34 PM, Katrin Fischer <katrin.fischer.83@web.de> wrote:

Hi Barton,

Control-number is the index on 001. 001 should have the number and 003 the MarcOrgCode, that's why it's stripped from $w for search. I don't know about OCLCs practices, so can't tell how numbers are handled there. The examples here show a number with ocm in 001:

http://www.loc.gov/marc/bibliographic/bd001.html

From that link:

Contains the control number assigned by the organization creating, using, or distributing the record. For interchange purposes, documentation of the structure of the control number and input conventions should be provided to exchange partners by the organization initiating the interchange.

That potentially means that we would have to write XSLT to transform the links in $w for each 'exchange partner' -- i.e. test the Marc Org Code, then apply a bunch of rules to generate a value that we can search for.

The examples don't leave me brimming with confidence that most exchange partners will use the same format for $w (after the Org Code) as for the 001:

001 #880524405##

003 CaOONL

001 ###86104385#

003 DLC

001 ocm14919759

003 OCoLC

001 #####9007496

003 DNLM

The description for $w (http://www.loc.gov/marc/bibliographic/bd76x78x.html) doesn't have a matching example:

"System control number of the related record preceded by the MARC code, enclosed in parentheses, for the agency to which the control number applies."

Hope this helps,

Katrin

Well, at the very least, it lets me know what I'm getting myself into.

I wonder if there's a way of translating the values found in $w into 001 outside of XSLT -- that's a language not well suited to the task. Could we do it in perl, and stash the results in some 9XX field?

I was kind of hoping that we would be able to use whatever we got back from extractControlNumber as a base for any transformations. That may or may not be a safe assumption.