Filtering XML records while bursting - processing only a subset of the data

Question

Hi
I'm attempting to filter the records in my XML data while bursting, so that depending on the value of a data element, the entire record is either skipped or included in the burst output.
I only want a subset of the XML output to be included in the burst output.
I've tried applying a filter at the select stage -
<xapi:request select="/AR_INVOICE/LIST_EMAIL_HEAD/EMAIL_HEAD{EMAIL_IND!=''}">
This removes all the records where 'EMAIL_IND' is not null - i.e. there is an email address to send to,
but instead of giving me multiple emails, one for each /AR_INVOICE/LIST_EMAIL_HEAD/EMAIL_HEAD,
I get just one mail address that contains all of the data for the records that have email addresses.
I also tried putting a filter on the template
<xapi:template type="rtf" locale="" location="xdo://AR.XXARINVEMAILDUMMY.en.00/?getSource=true" translation="" filter="{EMAIL_IND!=''}" />
i.e. having only one template and filtering as shown, but this has no effect.
Note: I had to change the square brackets - '[' to curly brackets '{' to get the examples to show.
Any ideas?
Thanks in advance.
Mike

Answer 1

Hi
I worked out a way to conditionally use ony a number of the data records in the bursting process - discarding the others.
In EBS, I generate a set of data in a a data-definition template that contains entries for some people who require email delivery, and some who require printed output.
In the concurrent programme, I specify an output rtf template that has a filter in it to print only the records for those who don't need emails.
(don't you just love the way that the designers call both the data definition and the output definition - TEPLATES - not confusing at all.....)
This step generates a sinlge sorted pdf file that is printed)
Then came the tricky part to send emails only to those who need them- using the bursting engine - to the other set of people - but to "cleanly" dispose of the records for the people who do not want emails.
What is not clear in any of the documentation at all is that the XMLPublisher bursting engine MUST handle ALL the records that it receives in the XML input.
If you specify a filter on the output template in the bursting control file, that excludes some records (those not needing the email) the bursting engine doesn't know what do with the remaining records.
Enter stage left - multiple delivery methods.
I simply defined another delivery method sample template with a filter including only those records for people who do NOT need emails - type filesystem - and routed the output to the /tmp directory - the trash.
The lesson(s) to be learnt
1) The bursting engine needs to have instructions to handle ALL the XML data that is fed to it.
2) You can define as many output documents and delivery methods as you like - putting filters on each delivery method as you like - BUT ALL XML RECORDS MUST be provided for.
3) XML records that are not required in one output / bursting stream can be handled in another - and trashed in the /tmp - or other - area.
The full bursting control file is shown below
1) First define the two delivery methods
2) Then define the two "documents" - using filters - using the two delivery methods.
Hope this helps others wanting to do similary things
Mike
<?xml version="1.0" encoding="UTF-8"?>
<xapi:requestset xmlns:xapi="http://xmlns.oracle.com/oxp/xapi" type="bursting">
<xapi:request select="/AR_INVOICE/LIST_EMAIL_HEAD/EMAIL_HEAD">
----- Define the two delivery methods - first the "email"---------
<xapi:delivery>
<xapi:email server="10.1.1.2" port="25"
from="${EM_DOC_EMAIL_ADDRESS}" reply-to ="${EM_COLLECTOR_EMAIL_ADDRESS}">
<xapi:message id="email" to="[email protected]"
attachment="true" subject="${EM_OPERATING_UNIT} invoice for ${EM_CUSTOMER_NAME}">
Please review the attached invoice - terms are ${EM_TERMS}
This will be sent to collector email ${EM_COLLECTOR_EMAIL_ADDRESS} and customer email ${EM_DOC_EMAIL_ADDRESS}
</xapi:message>
</xapi:email>
------ Second - the "null" - type filesystem ----------------
<xapi:filesystem id="null" output="/tmp/xmlp_null_${EM_CUSTOMER_NAME}" />
</xapi:delivery>
------Then define the first document using the"email" delivery -------------
<xapi:document output-type="pdf" delivery="email">
<xapi:template type="rtf" locale=""
location="xdo://AR.XXARINVEMAILDUMMY.en.00/?getSource=true" translation="" filter=".//EMAIL_HEAD[EM_DOC_EMAIL_ADDRESS!='NO_EMAIL']">
</xapi:template>
</xapi:document>
------ Then define the other document using the "null" delivery --------------
<xapi:document output-type="pdf" delivery="null">
     <xapi:template type="rtf" locale=""
     location="xdo://AR.XXARINVEMAILDUMMY.en.00/?getSource=true" translation="" filter=".//EMAIL_HEAD[EM_DOC_EMAIL_ADDRESS='NO_EMAIL']">
     </xapi:template>
</xapi:document>
</xapi:request>
</xapi:requestset>