Use of External tables to load XML data.

Question

Hi,
I have used external table definitions to load various XML files to the database, usually splitting the XML into seperate records - 1 per major element tag, and using PL/SQL to parse out a primary key to store in a relational table with all of the XML relevant to that primary key value stored as an XMLTYPE coumn in a row of the table. This has worked fine for XML with a single major entity (element tag) .
However, I now have an XML file that contains two "major" elements (both children of the root) that I would like to split out and store in seperate tables.
The XML file is of the following basic format:-
<drugs>
<drug>drug 1...</drug>
<drug>drug 2...</drug>
<partners>
<partner>partner 1</partner>
<partner>partner 2</partner>
</partners>
</drugs>
The problem is there are around 18000 elements of the first type, followed by several thousand of the 2nd type. I can create two seperate external tables - one for each element type, but how do I get the external table for the 2nd to ignore all the elements of the first type? My external table definition is :-
CREATE TABLE DRUGBANK_OWNER.DRUGBANK_PARTNERS_XML_EXTERNAL
DRUGBANK_XML CLOB
ORGANIZATION EXTERNAL
( TYPE ORACLE_LOADER
DEFAULT DIRECTORY DRUGBANK_DIR
ACCESS PARAMETERS
( records delimited by "</partner>" SKIP 100000
characterset al32utf8
badfile extlogs:'drugbank_partners_xml.bad'
logfile extlogs:'drugbank_partners_xml.log'
discardfile extlogs:'drugbank_partners_xml.dis'
READSIZE 52428800
fields
drugbank_xml CHAR(50000000) terminated by '</partners>'
LOCATION (DRUGBANK_DIR:'drugbank.xml')
REJECT LIMIT UNLIMITED
PARALLEL ( DEGREE 8 INSTANCES 1 )
NOMONITORING;
The problem is that before the first <partners> element the 1800 or so <drugs> elements cause a data cartrdige error:-
ORA-29913: error in executing ODCIEXTTABLEFETCH callout
ORA-29400: data cartridge error
KUP-04020: found record longer than buffer size supported, 52428800
This happens regardless of the value of the SKIP or the size of the drugbank_xml field.
I have tried using an OR on the "records delimited by" access parameter, to 'delimit by "</partner>" OR "</drug>"', with the intention of filtering out the <drug> elements but this leads to a syntax error.
Anyone ever tried anything similar and got it to work?
Any other suggestions?
Thanks,
Sid.

Answer 1

No, the content inside quotes is spanned across multiple lines....there are line breaks after every html tag.
"What's the error message you are getting?"
Iam not getting any error while selecting from external table , but I am getting tose rows in BAD file and log file has the following entries
KUP-04021: field formatting error for field TKBS_DSCN
KUP-04036: second enclosing delimiter not found
Message was edited by:
user627610