Document contains too many nodes error when extracting xml tag name

Question

I Have a large xml file in which the tag contains ~: as the value.
Now I am trying to extract all the tags which have ~: as the value and store that column using the following query and insert into a table.
insert into space_md select distinct xmltype(extract(value(x), '/').getstringval()).getrootelement() COLUMN_NAME
from gt_xmltype_tab gt, TABLE(XMLSequence(extract(gt.xmlfile1, '/ROWSET/ROW/*'))) x
where instr(extract(value(x),'/').getstringval(),'~:') > 1;
The xml file was generated using dbms_xmlgen.getxml. Table has 48 columns and around 4000 rows.
My above insert query gave me an error of 31186 too many nodes error.
I am using oracle version 10.2.0.3.
Following are the set of commands I ran....
SQL> insert into gt_xmltype_tab(xmlfile1)
values(XMLType(bfilename('BKUP_RES','QC.xml'),nls_charset_id('AL32UTF8'))); 2
1 row created.
SQL> SQL>
SQL> insert into restore_space_metadata select distinct 'QC', xmltype(extract(value(x), '/').getstringval()).getrootelement() COLUMN_NAME
2 from gt_xmltype_tab gt, TABLE(XMLSequence(extract(gt.xmlfile1, '/ROWSET/ROW/*'))) x
3 where instr(extract(value(x),'/').getstringval(),'~:') > 1;
insert into restore_space_metadata select distinct 'QC', xmltype(extract(value(x), '/').getstringval()).getrootelement() COLUMN_NAME
ERROR at line 1:
ORA-31186: Document contains too many nodes
Is there a better way of extracting the xml tag element name based on the xmltag content?
There is one other table which has 172 columns but only 100 rows so it doesnt create any problem on that table.
But this QC table has less columns but many many rows...
Any suggestions

Answer 1

There is a requiremnent of taking centain type of data backup and restore it.
It was implemented on flat file approach which was giving errors.
So it was implemented using XML approach.
Read data, store in xml file and read from xml file and load it into table.
Further, found that dbms_xmlstore is not able to handle tag only with whitespace
and tried to use the loading xml file into xmltype table column and extract data.
XMLTYPE column also has same problem of ignoring whitespace when used with extractvalue functions.
So for the workaround I update xmlfile having only one more more whitespace in the tag to have ~: character once.
After restoring data from xml to table I would run update qeury to update ~: to " ".
Now instead of running blind update for all the tables and all the columns from ~: to " " I thouhgt whyy not create a xml file of tag having ~:
along with its tablename.
and thats where I found the problem of too many nodes...
THe insert query you saw is populating table for table_name and column_name with tag having only ~: in it.
I hope this gives you the fair idea of stuff I am doing.