xml - XSLT Select non duplicate node values -

i trying non duplicate node values list. have tried many suggestions nothing quite worked me. source xml

<records>   <record>     <files>       <file>         <name>a</name>       </file>       <file>         <name>b</name>       </file>       <file>         <name>b</name>       </file>     </files>   </record>   <record>     <files>       <file>         <name>a</name>       </file>       <file>         <name>b</name>       </file>       <file>         <name>c</name>       </file>       <file>         <name>c</name>       </file>     </files>   </record> </records> 

output looking should looks text this

a,b | a,b,c 

first comma separated set first record , second set (after "|") second record. (delimiter placement, spaces etc... not issue, removal of duplicates)

code have looks this

<xsl:key name="nameid" match="name" use="." />  <xsl:for-each select="records/record">   <xsl:call-template name="doeach_record"/>   <xsl:text>|</xsl:text> </xsl:for-each>  <xsl:template name="doeach_record">   <xsl:for-each select="files/file">     <xsl:if test="generate-id(name) = generate-id(key('nameid', name)[1])">       <xsl:value-of select="name"/>     </xsl:if>     <xsl:text>,</xsl:text>   </xsl:for-each> </xsl:template> 

it remove duplicates first record, second record not selecting name values found in first record. getting output this

a,b | ,,c 

there xslt 2.0 since 2007 various implementations saxon 9, xmlprime, altova use <xsl:template match="record"><xsl:value-of select="distinct-values(files/file/name)" separator=","/></xsl:template>. if restricted xslt 1.0 need define key takes generated-id of ancestor record name e.g. <xsl:key name="unique" match="file/name" use="concat(generate-id(ancestor::record), '|', .)"/>, can use usual xslt 1.0 key based muenchian grouping/distinct value approach.