keys and performance

I am using XSLT to convert between two sets of values, using a "table"
defined in XML and a key.
Does the way the table is structured effect performance for large tables? 
For example, is either one of the following two structures better than the
other?
Or is there a better way?

Version 1: used elements will the keys stored as attributes
	<xsl:key name="CountryCodeKey" match="csl:CountryCode/csl:value"
use="@key"/>
	<csl:CountryCode default="Other">
		<csl:value key="IE">Ireland</csl:value>
		<csl:value key="FR">France</csl:value>
		<csl:value key="NL">Netherlands</csl:value>
		<csl:value key="JP">Japan</csl:value>
		<csl:value key="AU">Australia</csl:value>
	</csl:CountryCode >


Version 2: Uses separate elements for the keys and values
	<xsl:key name="CountryCodeKey" match="csl:CountryCode/csl:value"
use="preceding-sibling::cs:key[1]"/>
	<csl:CountryCode>
		<csl:key>IE</csl:key>
		<csl:value>Ireland</csl:value>
		<csl:key>FR</csl:key>
		<csl:value>France</csl:value>
		<csl:key>NL</csl:key>
		<csl:value>Netherlands</csl:value>
		<csl:key>JP</csl:key>
		<csl:value>Japan</csl:value>
		<csl:key>AU</csl:key>
		<csl:value>Australia</csl:value>
	</csl:CountryCode >

Thanks,

James Pasley

Cape Clear Software Limited.
61 Fitzwilliam Lane
Dublin 2
Ireland

Email : james.pasley@xxxxxxxxxxxxx
Switchboard  : + 353 (1) 2419900
Direct line  : + 353 (1) 2419957
Fax          : + 353 (1) 2419901
www          : www.capeclear.com



 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list