How to delete a node from Oracle XMLTYPE based on a condition?

Question

I have a XML data stored in a CLOB column and I would like to delete some nodes based on a specific condition.

Example XML Data :

<?xml version="1.0" encoding="UTF-8"?>
<payment>
  <person>
    <surname>Marco</surname>
    <name>Gralike</name>
    <salary>2345</salary>
  </person>
  <person>
    <surname>ABC</surname>
    <name>TEST</name>
    <salary>1234</salary>
    <person>
    <surname>Tiger</surname>
    <name>Scott</name>
    <salary>2222</salary>
  </person>
  </person>
 </payment>
 <payment>
  <person>
    <surname>BertJan</surname>
    <name>Meinders</name>
    <salary>3456</salary>
    <salary>125</salary>
  </person>
  <person>
    <surname>XYZ</surname>
    <name>TEST</name>
    <salary>1234</salary>
  </person>
 </payment>
 <payment>
  <person>
    <surname>Chris</surname>
    <name>Gralike</name>
    <salary>4567</salary>
  </person>
  <person>
    <surname>LMN</surname>
    <name>TEST</name>
    <salary>1234</salary>
  </person>
 </payment>

I need a Oracle PLSQL script to delete all the person tags if it contains TEST.

Final output would be :

<?xml version="1.0" encoding="UTF-8"?>
<payment>
  <person>
    <surname>Marco</surname>
    <name>Gralike</name>
    <salary>2345</salary>
  </person>
 </payment>
 <payment>
  <person>
    <surname>BertJan</surname>
    <name>Meinders</name>
    <salary>3456</salary>
    <salary>125</salary>
  </person>
 </payment>
 <payment>
  <person>
    <surname>Chris</surname>
    <name>Gralike</name>
    <salary>4567</salary>
  </person>
 </payment>

Thanks in advance.

Answer 1

Your provided XML doesn't have a root and can't be parsed by XML parser.

Assuming it does (say payments ) as shown below:

create table t(txt clob);
insert into t values('<?xml version="1.0" encoding="UTF-8"?>
<payments>
    <payment>
        <person>
            <surname>Marco</surname>
            <name>Gralike</name>
            <salary>2345</salary>
        </person>
        <person>
            <surname>ABC</surname>
            <name>TEST</name>
            <salary>1234</salary>
            <person>
                <surname>Tiger</surname>
                <name>Scott</name>
                <salary>2222</salary>
            </person>
        </person>
    </payment>
    <payment>
        <person>
            <surname>BertJan</surname>
            <name>Meinders</name>
            <salary>3456</salary>
            <salary>125</salary>
        </person>
        <person>
            <surname>XYZ</surname>
            <name>TEST</name>
            <salary>1234</salary>
        </person>
    </payment>
    <payment>
        <person>
            <surname>Chris</surname>
            <name>Gralike</name>
            <salary>4567</salary>
        </person>
        <person>
            <surname>LMN</surname>
            <name>TEST</name>
            <salary>1234</salary>
        </person>
    </payment>
</payments>');

You can use this:

update t
set txt = to_clob(deletexml(
  xmltype(t.txt),
  '//payment/person[./name[text()="TEST"]]'
));

Produces:

<?xml version="1.0" encoding="UTF-8"?>
<payments>
    <payment>
        <person>
            <surname>Marco</surname>
            <name>Gralike</name>
            <salary>2345</salary>
        </person>
    </payment>
    <payment>
        <person>
            <surname>BertJan</surname>
            <name>Meinders</name>
            <salary>3456</salary>
            <salary>125</salary>
        </person>
    </payment>
    <payment>
        <person>
            <surname>Chris</surname>
            <name>Gralike</name>
            <salary>4567</salary>
        </person>
    </payment>
</payments>

EDIT:

If you want to delete a node that doesn't have a given child, use this:

update t
set txt = to_clob(deletexml(
  xmltype(t.txt),
  '//payment[not(./person)]'
));

It'll delete all the payment tags that don't have a person in it.

Answer 2

deleteXML() has been deprecated. If possible you should use XQuery update. Also try to avoid the use of '//' if the full path is fixed.

with XML_TABLE as 
(
   select XMLTYPE('<?xml version="1.0" encoding="UTF-8"?>
<payments>
<payment>
    <person>
        <surname>Marco</surname>
        <name>Gralike</name>
        <salary>2345</salary>
    </person>
    <person>
        <surname>ABC</surname>
        <name>TEST</name>
        <salary>1234</salary>
        <person>
            <surname>Tiger</surname>
            <name>Scott</name>
            <salary>2222</salary>
        </person>
    </person>
</payment>
<payment>
    <person>
        <surname>BertJan</surname>
        <name>Meinders</name>
        <salary>3456</salary>
        <salary>125</salary>
    </person>
    <person>
        <surname>XYZ</surname>
        <name>TEST</name>
        <salary>1234</salary>
    </person>
</payment>
<payment>
    <person>
        <surname>Chris</surname>
        <name>Gralike</name>
        <salary>4567</salary>
    </person>
    <person>
        <surname>LMN</surname>
        <name>TEST</name>
        <salary>1234</salary>
    </person>
</payment>
</payments>') as XML_COLUMN from dual
)
SELECT XMLQuery(
    'copy $NEWXML := $XML modify (
      delete nodes $NEWXML/payments/payment/person[name[text()=$NAME]]
     )
     return $NEWXML'
     passing XML_COLUMN as "XML",
             'TEST' as "NAME"
     returning CONTENT
   )
from XML_TABLE
/

You can try this code snipped using the SQL Workbench at livesql.oracle.com

Answer 3

Thanks for the sample.

<payment> 
<person>
 <surname>XYZ</surname> 
 <name>TEST</name> 
 <salary>1234</salary> 
 </person> 
<id>person 4</id> 
</payment> 

If one of the node contains above data then after using deletexml i get this :

<payment> 
<id>person 4</id> 
</payment> 

How to delete this node if it does not contain any <person> tag ?

ie How to delete the below Node as it does not contain <person> tag :

   <payment> 
    <id>person 4</id> 
    </payment>

Answer 4

This works

with XML_TABLE as
(
   select XMLTYPE('<?xml version="1.0" encoding="UTF-8"?>
<payments>
  <payment>
    <person>
     <surname>XYZ</surname>
  <name>TEST</name>
  <salary>1234</salary>
</person>
<id>person 4</id>
 </payment>
 <payment>
<id>person 5</id>
</payment>
 </payments>') as XML_COLUMN from dual
)
SELECT XMLQuery(
'copy $NEWXML := $XML modify (
  delete nodes $NEWXML/payments/payment[not(person)]
 )
 return $NEWXML'
 passing XML_COLUMN as "XML",
         'TEST' as "NAME"
 returning CONTENT
)
from XML_TABLE

<?xml version="1.0" encoding="WINDOWS-1252"?>
<payments>
  <payment>
 <person>
  <surname>XYZ</surname>
  <name>TEST</name>
  <salary>1234</salary>
</person>
  <id>person 4</id>
 </payment>
</payments>

SQL>