?xml version="1.0"? tag not appearing as first characters in document

Question

Hi,
JSP below successfully creates a XML document but it includes a blank line before the <?xml version="1.0"?> tag.
This causes my PL/SQL to return a "ORA-20100: Error occurred while parsing: PI names starting with 'xml' are
reserved." error when using the XMLPARSER package.
I am outputting the XML to IE5.0 but even if I do a SYSTEM out I get the same blank line before the initial tag.
There are posts on here that confirm the PI error is caused by the tag not being the first characters in the document, but no solution/fix is provided.
Any ideas much appreciated.
JSP Code
<%@ page import="java.sql.*, oracle.jbo.*, oracle.jdeveloper.cm.*, oracle.jdbc.*,oracle.xml.sql.query.*" %>
<%String driver="oracle.jdbc.driver.OracleDriver";
Driver d = new oracle.jdbc.driver.OracleDriver();
String dbURL="jdbc:oracle:thin:@localhost:1521:mydb";
String login="i2k";
String password="fred";
Connection cn = null;
cn = DriverManager.getConnection(dbURL,login,password);
// SQL Statement from URL Parameters
String sql = request.getParameter("sql");
if(sql == null){
sql = "select * from vfi_trans";
// Create SQL-to-XML Handler
OracleXMLQuery q = new OracleXMLQuery(cn, sql);
// Use <TransactionList> as document element for Rowset
q.setRowsetTag("TransactionList");
// Use <Transaction> for each row in the result
q.setRowTag("Transaction");
// set encoding
q.setEncoding("iso-8859-1");
// ensure lower case element names
q.useLowerCaseTagNames();
// Generate XML results and write to output
String xmldoc = q.getXMLString();
out.println(xmldoc.trim());
//System.out.println(xmldoc.indexOf("\n"));
cn.close();%>
PL/SQL
PROCEDURE XML_HANDLER2 IS
-- MODIFICATION HISTORY
-- Person Date Comments
vfiURL VARCHAR2(100);
parser xmlparser.Parser;
vfiXML xmldom.DOMDocument;
transactions xmldom.DOMNodeList;
transactions_found NUMBER;
curNode xmldom.DOMNode;
textChild xmldom.DOMNode;
v_itrans_site vfi_trans.itrans_site%TYPE;
BEGIN
dbms_output.put_line('Integrator 2000 Transactions');
-- This is the URL to browse for an XML-based vfi feed of stories on XML
vfiURL := 'http://10.1.1.111:7070/i2k25_html/ShowQuery.jsp?sql=select%20*%20from%20vfi_trans';
-- Set the machine to use as the HTTP proxy server for URL requests
http_util.setProxy('MYPROXY');
-- Parse the live XML vfi feed from Moreover.com by URL
parser := xmlparser.newParser;
vfiXML := xml.parseURL( vfiURL );
xmlparser.freeParser(parser);
-- Search for all <headline_text> elements in the document we recieve
transactions := xpath.selectNodes(vfiXML,'/TransactionList/ITRANS_ID');
-- Loop over the "hits" and print out the text of the title
FOR j IN 1..xmldom.getLength(transactions) LOOP
-- Get the current <headline_text> node (Note the list is zero-based!)
curNode := xmldom.item(transactions,j-1);
-- The text of the title is the first child (text) node of
-- the <headline_text> element in the list of "hits"
-- textChild := xmldom.getFirstChild(curNode);
v_itrans_site := xpath.valueof(curNode, '.');
dbms_output.put_line('('| |LPAD(j,2)| |') '| | v_itrans_site);
END LOOP;
-- Free the XML document full of vfi stories since we're done with it.
xml.freeDocument(vfiXML);
EXCEPTION
WHEN OTHERS THEN
RAISE;
END; -- Procedure
null

Answer 1

Charles,
I believe that the blank line is caused by the JSP engine when it strips out the '<%@ page import...>' (replace bracket with brace) statement. God (or at least Larry E) forgive me for posting a link at IBM, but this article speaks to your issue:
http://www-106.ibm.com/developerworks/library/j-dynxml.html?dwzone=ibm
Maybe you need to put the <?xml?> tag in the jsp itself and strip it out of your xmldoc before outputting it.
Good luck.