java.lang.OutOfMemoryError only with XSLTC on 5MB document

[Jeff Baker <JB...@ehsmed.com>]

Hi -

I am transforming an XML document usually between 4-5MB in size. When
transforming using Xalan (no XSLTC), everything works great. However, when
setting XSLTC on, I always get a java.lang.OutOfMemoryError.

Here are some facts about the stylesheet. I am converting a relatively flat
structure to a hierarchical structure and using the Muenchian method. I am
grouping around 1800-2000 nodes. I am heavily using keys. I am also heavily
using <xsl:copy-of>. I am posting on dev (regarding the Out of Memory
Error), and on users (for recommendations for improving the stylesheet).

I have included the stylesheet below:

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

<!--
    Document   : processClaims.xsl
    Created on : September 30, 2002, 6:45 PM
    Author     : jbaker
    Description:
        Purpose of transformation follows.
-->

<xsl:stylesheet version="1.0" 
  xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
  
  <!-- set output mode as XML -->  
  <xsl:output method="xml" />
  
  <!-- define the key indexes -->
  <!-- index claims using the Muenchian method for grouping -->
  <!-- index claims by unique Financial Group and Provider -->
  <xsl:key name="pf" match="/rootElement/claims/claim"
    use="concat(@financialGroupId,'.',providerInfo/@renderingProviderId)" />
  <!-- index claims by unique Financial Group AND Provider AND SUBSCRIBER
-->
  <xsl:key name="pf_s" match="/rootElement/claims/claim"
    use="concat(@financialGroupId,'.',providerInfo/@renderingProviderId,
        '.',subscribers/subscriberInfo[1]/@subscriberId)" />
  <!-- index claims by unique Financial Group AND Provider AND Subscriber 
       AND Patient (where subscriber is not the patient) -->
  <xsl:key name="pf_s_p"
match="/rootElement/claims/claim[not(subscribers/subscriberInfo[1]/@patientS
ubscriberRelation='SELF')]"
    use="concat(@financialGroupId,'.',providerInfo/@renderingProviderId,
        '.',subscribers/subscriberInfo[1]/@subscriberId,'.',@patientId)" />

  <!-- index providers by id -->
  <xsl:key name="pv" match="/rootElement/providers/provider" use="@id" />
  <!-- index financial groups by id -->
  <xsl:key name="fg" match="/rootElement/financialGroups/financialGroup"
use="@id" />
  <!-- index subscribers by id -->
  <xsl:key name="sb" match="/rootElement/subscribers/subscriber" use="@id"
/>
  <!-- index payers by id -->
  <xsl:key name="py" match="/rootElement/payers/payer" use="@id" />
  <!-- index patients by id -->
  <xsl:key name="pt" match="/rootElement/patients/patient" use="@id" />
  <!-- index service locations by id -->
  <xsl:key name="sl" match="/rootElement/serviceLocations" use="@id" />
  <!-- index service line items by claim id -->
  <xsl:key name="sli" match="/rootElement/serviceLines/serviceLine"
use="@patientVisitId" />
  
  <!-- template rule matching source root element -->
  <xsl:template match="/">
    <xsl:element name="claimFile">
      <!-- loop over the distinct provider/financial group combinations -->
      <xsl:for-each select="/rootElement/claims/claim[generate-id(.) = 
 
generate-id(key('pf',concat(@financialGroupId,'.',providerInfo/@renderingPro
viderId)))]">
        <!-- with each provider/financial group, call
processBillingInformation -->
        <xsl:call-template name="processBillingInformation">
          <xsl:with-param name="financialGroupId" select="@financialGroupId"
/>
          <xsl:with-param name="providerId"
select="providerInfo/@renderingProviderId" />
        </xsl:call-template>
      </xsl:for-each>
    </xsl:element>
  </xsl:template>

  <!-- processBillingInformation closely models LOOP2000A in HIPAA 837 -->
  <xsl:template name="processBillingInformation">
    <xsl:param name="financialGroupId" />
    <xsl:param name="providerId" />    
    <xsl:element name="billingInformation">
      <xsl:copy-of select="key('fg',$financialGroupId)" />
      <xsl:copy-of select="key('pv',$providerId)" />
      <!-- query of all the claims for the current financial group/provider
-->        
      <xsl:variable name="pf_query" 
        select="key('pf',concat($financialGroupId,'.',$providerId))" />

      <!-- loop over the distinct subscribers in each financial group -->
      <xsl:for-each select="$pf_query[generate-id(.) = 
        generate-id(key('pf_s',
concat($financialGroupId,'.',$providerId,'.',
        subscribers/subscriberInfo[1]/@subscriberId)))]">        
        <xsl:call-template name="processSubscriberInformation">
          <xsl:with-param name="financialGroupId" select="$financialGroupId"
/>
          <xsl:with-param name="providerId" select="$providerId" />
          <xsl:with-param name="subscriberId"
select="subscribers/subscriberInfo[1]/@subscriberId" />
        </xsl:call-template>        
      </xsl:for-each>
    </xsl:element>
  </xsl:template>
  
  <!-- processSubscriberInformation closely models LOOP2000B in HIPAA 837
-->
  <xsl:template name="processSubscriberInformation">
    <xsl:param name="financialGroupId" />
    <xsl:param name="providerId" />
    <xsl:param name="subscriberId" />    
    <xsl:element name="subscriberInformation">
      <xsl:variable name="payerId" select="key('sb',$subscriberId)/@payerId"
/>
      <xsl:copy-of select="key('sb',$subscriberId)" />
      <xsl:copy-of select="key('py',$payerId)" />      
      <!-- retrieve claims related to subscriber where subscriber is the
patient -->
      <xsl:variable name="pf_s_query"
        select="key('pf_s',concat($financialGroupId,'.',
 
$providerId,'.',$subscriberId))[subscribers/subscriberInfo[1]/@patientSubscr
iberRelation='SELF']" />
      <!-- for each claim, call processClaimInformation -->
      <xsl:for-each select="$pf_s_query">
        <xsl:call-template name="processClaimInformation">
          <xsl:with-param name="claim" select="." />
        </xsl:call-template>
      </xsl:for-each>      
      <!-- retrieve claims related to subscriber where subscriber is NOT the
patient -->
      <xsl:variable name="pf_s_query2"
select="key('pf_s',concat($financialGroupId,'.',
 
$providerId,'.',$subscriberId))[not(subscribers/subscriberInfo[1]/@patientSu
bscriberRelation='SELF')]" />
      <!-- loop over the distinct patients from claims and call
processPatientInformation -->
      <xsl:for-each select="$pf_s_query2[generate-id(.) =
generate-id(key('pf_s_p',
 
concat($financialGroupId,'.',$providerId,'.',$subscriberId,'.',@patientId)))
]">        
        <xsl:call-template name="processPatientInformation">
          <xsl:with-param name="financialGroupId" select="$financialGroupId"
/>
          <xsl:with-param name="providerId" select="$providerId" />
          <xsl:with-param name="subscriberId" select="$subscriberId" />
          <xsl:with-param name="patientId" select="@patientId" />
        </xsl:call-template>        
      </xsl:for-each>
    </xsl:element>  
  </xsl:template>
  
  <!-- processPatientInformation closely models LOOP2000C in HIPAA 837 -->
  <xsl:template name="processPatientInformation">
    <xsl:param name="financialGroupId" />
    <xsl:param name="providerId" />
    <xsl:param name="subscriberId" />
    <xsl:param name="patientId" />    
    <xsl:element name="patientInformation">
      <xsl:copy-of select="key('pt',$patientId)" />
      <xsl:variable name="pf_s_p_query"
select="key('pf_s_p',concat($financialGroupId,'.',
        $providerId,'.',$subscriberId,'.',$patientId))" />      
      <xsl:for-each select="$pf_s_p_query">
        <xsl:call-template name="processClaimInformation">
          <xsl:with-param name="claim" select="." />
        </xsl:call-template>
      </xsl:for-each>              
    </xsl:element>    
  </xsl:template>
  
  <!-- processClaimInformation closely models LOOP2300 in HIPAA 837 -->
  <xsl:template name="processClaimInformation">
    <xsl:param name="claim" />
    <xsl:element name="claimInformation">
      <xsl:copy-of select="$claim" />
      <!-- insert all the service line items associated with the claim.
These
      closely model LOOP2400 in HIPAA 837 -->
      <xsl:copy-of select="key('sli',$claim/@id)" />
    </xsl:element>
  </xsl:template>
  
</xsl:stylesheet> 

Re: java.lang.OutOfMemoryError only with XSLTC on 5MB document

[Santiago Pericas-Geertsen <Sa...@sun.com>]

On Tuesday 01 October 2002 18:23, Jeff Baker wrote:
> Hi -
>
> I am transforming an XML document usually between 4-5MB in size. When
> transforming using Xalan (no XSLTC), everything works great. However, when
> setting XSLTC on, I always get a java.lang.OutOfMemoryError.
>

 Jeff,

 Two questions: What version of XSLTC are you using? (try  "java 
org.apache.xalan.xsltc.cmdline.Compile -v") Could you provide an XML document 
too?

 Thanks.

-- Santiago