Optimize XML query

Question

Hi,
I'm on 10.2.0.3 and got query like that:
desc SSS_ARCHIVE2
Name                                      Null?    Type
SSS_TASK_ID                                        VARCHAR2(32)
SSS_ID                                             VARCHAR2(30)
SSS_CORRELATION_ID                                 VARCHAR2(32)
SSS_SKELETON_ID                                    VARCHAR2(32)
SSS_CUSTOMER_ID                                    VARCHAR2(32)
SSS_OPERATOR                                       VARCHAR2(32)
SSS_STATUS_DATE                           NOT NULL DATE
SSS_AUDIT_UC                              NOT NULL VARCHAR2(512)
SSS_AUDIT_DC                              NOT NULL TIMESTAMP(6)
SSS_AUDIT_UM                                       VARCHAR2(512)
SSS_AUDIT_DM                                       TIMESTAMP(6)
SSS_AUDIT_ST                              NOT NULL NUMBER(1)
SSS_CREATION_DATE                                  DATE
SSS_DESCRIPTION                                    VARCHAR2(512)
SSS_DATA                                           BLOB
SSS_DATA2                                          CLOB
create table ARH_2 as
SELECT        SSS_task_id,  extractValue(value(taskhist),'/task/taskId') taskId,
              extractValue(value(taskhist),  '/task/id') id_,
              extractValue(value(typeId),  '/typeId') typeId,
              rownum wiersz FROM
              sss_archive2,     
              TABLE(xmlsequence(extract(xmltype(SSS_archive2.SSS_data2),'/ns2:ArchivisedTask/taskHistory/task','xmlns:ns2="http://some.com/qservice/model" xmlns:ns3="http://some.com/qservice/method"'))) taskhist,
              TABLE (xmlsequence(extract(value(taskhist),'/task/typeId'))) typeId
              where
                SSS_task_id=extractValue(value(taskhist),  '/task/taskId')  
                and SSS_audit_st=1
Plan hash value: 2424165471
| Id  | Operation                             | Name                   | E-Rows |  OMem |  1Mem | Used-Mem |
|   1 |  LOAD AS SELECT                       |                        |        |   519K|   519K|  519K (0)|
|   2 |   COUNT                               |                        |        |       |       |          |
|   3 |    NESTED LOOPS                       |                        |    267M|       |       |          |
|   4 |     NESTED LOOPS                      |                        |  16360 |       |       |          |
|*  5 |      TABLE ACCESS FULL                | SSS_ARCHIVE2           |    631K|       |       |          |
|*  6 |      COLLECTION ITERATOR PICKLER FETCH| XMLSEQUENCEFROMXMLTYPE |        |       |       |          |
|   7 |     COLLECTION ITERATOR PICKLER FETCH | XMLSEQUENCEFROMXMLTYPE |        |       |       |          |
Predicate Information (identified by operation id):
   5 - filter("SSS_AUDIT_ST"=1)
   6 - filter("SSS_TASK_ID"=EXTRACTVALUE(VALUE(KOKBF$),'/task/taskId'))
Note
   - dynamic sampling used for this statement
   - Warning: basic plan statistics not available. These are only collected when:
       * hint 'gather_plan_statistics' is used for the statement or
       * parameter 'statistics_level' is set to 'ALL', at session or system levelThere is about 582966 records in SSS_ARCHIVE2 and SSS_AUDIT_ST=1 is none selective 90% rows.
Is there a proper way to extract data from xml ?
Regards.
Greg

Answer 1

Ah, yes the namespaces can be a pain (blame Microsoft, it was their idea).
Here's another example using namespaces...
-- With namespace
SQL> ed
Wrote file afiedt.buf
  1  WITH t as (select XMLTYPE('
  2  <RECSET xmlns:aa="http://www.w3.org">
  3    <aa:REC>
  4      <aa:COUNTRY>1</aa:COUNTRY>
  5      <aa:POINT>1800</aa:POINT>
  6      <aa:USER_INFO>
  7        <aa:USER_ID>1</aa:USER_ID>
  8        <aa:TARGET>28</aa:TARGET>
  9        <aa:STATE>6</aa:STATE>
10        <aa:TASK>12</aa:TASK>
11      </aa:USER_INFO>
12      <aa:USER_INFO>
13        <aa:USER_ID>5</aa:USER_ID>
14        <aa:TARGET>19</aa:TARGET>
15        <aa:STATE>1</aa:STATE>
16        <aa:TASK>90</aa:TASK>
17      </aa:USER_INFO>
18    </aa:REC>
19    <aa:REC>
20      <aa:COUNTRY>2</aa:COUNTRY>
21      <aa:POINT>2400</aa:POINT>
22      <aa:USER_INFO>
23        <aa:USER_ID>3</aa:USER_ID>
24        <aa:TARGET>14</aa:TARGET>
25        <aa:STATE>7</aa:STATE>
26        <aa:TASK>5</aa:TASK>
27      </aa:USER_INFO>
28    </aa:REC>
29  </RECSET>') as xml from dual)
30  -- END OF TEST DATA
31  select x.country, x.point, y.user_id, y.target, y.state, y.task
32  from t
33      ,XMLTABLE(XMLNAMESPACES('http://www.w3.org' as "aa"),
34                '/RECSET/aa:REC'
35                PASSING t.xml
36                COLUMNS country NUMBER PATH '/aa:REC/aa:COUNTRY'
37                       ,point   NUMBER PATH '/aa:REC/aa:POINT'
38                       ,user_info XMLTYPE PATH '/aa:REC/*'
39               ) x
40      ,XMLTABLE(XMLNAMESPACES('http://www.w3.org' as "aa"),
41                '/aa:USER_INFO'
42                PASSING x.user_info
43                COLUMNS user_id NUMBER PATH '/aa:USER_INFO/aa:USER_ID'
44                       ,target  NUMBER PATH '/aa:USER_INFO/aa:TARGET'
45                       ,state   NUMBER PATH '/aa:USER_INFO/aa:STATE'
46                       ,task    NUMBER PATH '/aa:USER_INFO/aa:TASK'
47*              ) y
SQL> /
   COUNTRY      POINT    USER_ID     TARGET      STATE       TASK
         1       1800          1         28          6         12
         1       1800          5         19          1         90
         2       2400          3         14          7          5
SQL>