Multiple Level Grouping using something like position(

> Hi,
> 
> I have been beating my head against the wall on something, I was hoping to get help with.
> 
> I have an XML File
> <?xml version="1.0" encoding="UTF-8"?>
> <?xml-stylesheet type="text/xsl" href="S:\Database\submissions\xslt_reports\subject_list.xsl"?>
> <pharmgkb xmlns="http://www.pharmgkb.org/schema/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.pharmgkb.org/schema/ http://www.pharmgkb.org/schema/root.xsd">
> 	<gene pharmgkbId="PA117">
> 		<referenceSequence>
> 			<dnaSequence>ccagTAAGCGCCCTCCTAATCCCCGCAGCGCCACC</dnaSequence>
> 			<experiment>
> 				<sampleSetXref resource="PharmGKB">PA128747821</sampleSetXref>
> 				<genotypesInSubject>
> 					<subjectXref resource="PharmGKB">PA126746586</subjectXref>
> 					<pcrResult>
> 						<assayXref resource="local">18</assayXref>
> 						<sequencedBothStrands>true</sequencedBothStrands>
> 					</pcrResult>
> 				</genotypesInSubject>
> 				<genotypesInSubject>
> 					<subjectXref resource="PharmGKB">PA126746600</subjectXref>
> 					<pcrResult>
> 						<assayXref resource="local">18</assayXref>
> 						<sequencedBothStrands>true</sequencedBothStrands>
> 					</pcrResult>
> 				</genotypesInSubject>
> 				<genotypesInSubject>
> 					<subjectXref resource="PharmGKB">PA126746573</subjectXref>
> 					<pcrResult>
> 						<assayXref resource="local">18</assayXref>
> 						<sequencedBothStrands>true</sequencedBothStrands>
> 					</pcrResult>
> 				</genotypesInSubject>
> 			</experiment>
> 		</referenceSequence>
> 		<referenceSequence>
> 			<dnaSequence>cttcTGCTGTCTCTTCTGAG</dnaSequence>
> 			<experiment>
> 				<sampleSetXref resource="PharmGKB">PA128747821</sampleSetXref>
> 				<genotypesInSubject>
> 					<subjectXref resource="PharmGKB">PA126746541</subjectXref>
> 					<pcrResult>
> 						<assayXref resource="local">19</assayXref>
> 						<sequencedBothStrands>true</sequencedBothStrands>
> 						<variant>
> 							<position>320</position>
> 							<firstAllele>G</firstAllele>
> 							<secondAllele>G</secondAllele>
> 						</variant>
> 						<variant>
> 							<position>315</position>
> 							<firstAllele>C</firstAllele>
> 							<secondAllele>G</secondAllele>
> 						</variant>
> 						<variant>
> 							<position>465</position>
> 							<firstAllele>T</firstAllele>
> 							<secondAllele>C</secondAllele>
> 						</variant>
> 						<variant>
> 							<position>194</position>
> 							<firstAllele>C</firstAllele>
> 							<secondAllele>C</secondAllele>
> 						</variant>
> 					</pcrResult>
> 				</genotypesInSubject>
> 				<genotypesInSubject>
> 					<subjectXref resource="PharmGKB">PA126746573</subjectXref>
> 					<pcrResult>
> 						<assayXref resource="local">19</assayXref>
> 						<sequencedBothStrands>true</sequencedBothStrands>
> 						<variant>
> 							<position>194</position>
> 							<firstAllele>C</firstAllele>
> 							<secondAllele>C</secondAllele>
> 						</variant>
> 						<variant>
> 							<position>465</position>
> 							<firstAllele>T</firstAllele>
> 							<secondAllele>C</secondAllele>
> 						</variant>
> 						<variant>
> 							<position>315</position>
> 							<firstAllele>C</firstAllele>
> 							<secondAllele>G</secondAllele>
> 						</variant>
> 						<variant>
> 							<position>320</position>
> 							<firstAllele>G</firstAllele>
> 							<secondAllele>G</secondAllele>
> 						</variant>
> 					</pcrResult>
> 				</genotypesInSubject>
> 				<genotypesInSubject>
> 					<subjectXref resource="PharmGKB">PA126746574</subjectXref>
> 					<pcrResult>
> 						<assayXref resource="local">19</assayXref>
> 						<sequencedBothStrands>true</sequencedBothStrands>
> 						<variant>
> 							<position>320</position>
> 							<firstAllele>G</firstAllele>
> 							<secondAllele>G</secondAllele>
> 						</variant>
> 						<variant>
> 							<position>315</position>
> 							<firstAllele>C</firstAllele>
> 							<secondAllele>G</secondAllele>
> 						</variant>
> 						<variant> >
> 							<position>194</position>
> 							<firstAllele>C</firstAllele>
> 							<secondAllele>C</secondAllele>
> 						</variant>
> 						<variant>
> 							<position>465</position>
> 							<firstAllele>T</firstAllele>
> 							<secondAllele>C</secondAllele>
> 						</variant>
> 					</pcrResult>
> 				</genotypesInSubject>
> 				<genotypesInSubject>
> 					<subjectXref resource="PharmGKB">PA126746575</subjectXref>
> 					<pcrResult>
> 						<assayXref resource="local">19</assayXref>
> 						<sequencedBothStrands>true</sequencedBothStrands>
> 						<variant>
> 							<position>465</position>
> 							<firstAllele>C</firstAllele>
> 							<secondAllele>C</secondAllele>
> 						</variant>
> 						<variant>
> 							<position>315</position>
> 							<firstAllele>G</firstAllele>
> 							<secondAllele>G</secondAllele>
> 						</variant>
> 						<variant>
> 							<position>194</position>
> 							<firstAllele>C</firstAllele>
> 							<secondAllele>C</secondAllele>
> 						</variant>
> 						<variant>
> 							<position>320</position>
> 							<firstAllele>G</firstAllele>
> 							<secondAllele>G</secondAllele>
> 						</variant>
> 					</pcrResult>
> 				</genotypesInSubject>
> 			</experiment>
> 		</referenceSequence>
> 	</gene>
> </pharmgkb>
> 
> I need to take this and make it into a spreadsheet like
> 
>                               Experiment 1
> subject      Variant1 Variant2 Variant3
> 1			A/A	         A/G
> 2                  C/A
> 3
> 
> So I thought I would output it as csv file like (excluding experiments for now)
> ,Variant1,Variant2,...
> subjectID,allele1/allele2,etc.
> 
What I need then is the input doc, turned to be by subject, then for each subject, output a row of data, grouped by experiment, then variant position.

> I have tried many different things with keys, variables, muenchian grouping, and the best I can do so far is get the subJectXref   (unique subjects)  .  I have to this point been very unsuccessful making layers of grouping.  I thought I could make a key for an <experiment> node using position, but I found a post on how position() won't work in  key, since it will always be 1.  
> 
> I haven't worked with XSLT for a few years now, and seem to be having trouble getting anything that works.
> 
<xsl:stylesheet version="1.0" xmlns:xsl="http://www.w3.org/1999/XSL/Transform" xmlns:n1="http://www.pharmgkb.org/schema/" xmlns:sch="http://www.ascc.net/xml/schematron" xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
	<xsl:output method="text" indent="yes"/>
	<!--index all the subject elements in doc by value of subjectXref-->
	<xsl:key name="subject-key" match="//n1:subjectXref" use="."/>
	<xsl:key name="exp-variant-key" match="//n1:variant" use="n1:position"/>
	<xsl:template match="/">
		<!--get unique subjects -->
		<xsl:variable name="unique-subjects" select="//n1:subjectXref[generate-id(.)=generate-id(key('subject-key', .))]"/>
		<!--get all the experiments-->
		<xsl:variable name="expNodes" select="//n1:experiment"/>
		<!--get all the variants-->
		<xsl:variable name="varNodes" select="//n1:variant"/>
		<!--loop on unique subjects-->
		<xsl:for-each select="$unique-subjects">
			<!-- Save off subject value -->
			<xsl:variable name="subj" select="."/>
			<!-- Output carriage return, subjectID, which is the value of the element subjectXref, and comma-->
			<xsl:text/>
			<xsl:value-of select="."/>,
			<!-- for each experiment -->
			<xsl:for-each select="$expNodes">
				<!--get all genotypes in each experient for the current subject-->
				<xsl:variable name="genotypeNode" select="n1:genotypesInSubject[n1:subjectXref=$subj]"/>
				<!--get unique variants by experiment-->
				<xsl:variable name="unique-exp-variants" select="//n1:variant[generate-id(.)=generate-id(key('exp-variant-key',n1:position))]"/>
				<xsl:for-each select="$unique-exp-variants">
					<xsl:if test="../../../[n1:experient=$unique-exp-variants]">
						<xsl:value-of select="./n1:position"/>,
					</xsl:if>
				</xsl:for-each>
				<!--get all unique variants for each experiment-->
				<xsl:for-each select="n1:pcrResult/n1:variant">
					<allele1>
						<xsl:value-of select="n1:allele1"/>
					</allele1>
					<allele2>
						<xsl:value-of select="n1:allele2"/>
					</allele2>
				</xsl:for-each>
			</xsl:for-each>
		</xsl:for-each>
	</xsl:template>
</xsl:stylesheet>

Thanks!

-mat  

> I have looked at the info from Jeni's site on grouping, and multiple grouping, and trolled for other posts, but have yet to translate that to working bits of code.  And I am doing this in XML Spy with default processor, if that helps.
> 
> Anyway, if anyone has some quick pointers (or some extra brains they can send for me to insert into my head) that would be nice :-)
> 
> 
> 
> -mat
> 
> 
> 

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