假设我有一个表格如下结构：
创建表Temp（ID数字，数据xmltype）;
并且标识为1的内容为
<PivotSet>
<item>
<column name = "ATTRIBUTENAME">test</column>
<column name = "MAX(ATTRIBUTEVALUE)">testVal</column>
</item>
<item>
<column name = "ATTRIBUTENAME">test1</column>
<column name = "MAX(ATTRIBUTEVALUE)">test1Val</column>
</item>
<item>
<column name = "ATTRIBUTENAME">test2</column>
<column name = "MAX(ATTRIBUTEVALUE)">test2Val</column>
</item>
现在我一直渴望实现的是将xml中的元素展平，以得到这个结果：
id | test | test 1 | test 2 |
1 |testval |testval1 |testval2 |
所以我想要的是显示的每个项目都有第一个<column>为columnName，第二个作为该列的值。
答案 0 :(得分：1)
SQL> select * from t;
ID
----------
DATA
--------------------------------------------------------------------------------
1
<PivotSet>
<item>
<column name="ATTRIBUTENAME">test</column>
<column name="MAX(ATTRIBUTEVALUE)">testVal</column>
</item>
<item>
<column name="ATTRIBUTENAME">test1</column>
<column name="MAX(ATTRIBUTEVALUE)">test1Val</column>
</item>
<item>
<column name="ATTRIBUTENAME">test2</column>
<column name="MAX(ATTRIBUTEVALUE)">test2Val</column>
</item>
</PivotSet>
SQL> select 'ID' cn,
2 max(decode(rn,1,column_name,null)) a,
3 max(decode(rn,2,column_name,null)) b,
4 max(decode(rn,3,column_name,null)) c,
5 id,
6 max(decode(rn,1,value,null)) d,
7 max(decode(rn,2,value,null)) e,
8 max(decode(rn,3,value,null)) f
9 from (
10 select id, extractValue(value(r),'item/column[@name = "ATTRIBUTENAME"]') column_name
11 , extractValue(value(r),'item/column[@name = "MAX(ATTRIBUTEVALUE)"]') value
12 , row_number() over(partition by id order by null) rn
13 from t,
14 table(XMLSequence((t.data).extract('PivotSet/item'))) r
15 ) group by id
16 /
CN A B C ID D E F
-- -------- -------- -------- ---- -------- -------- --------
ID test test1 test2 1 testVal test1Val test2Val
或（对于新版本的问题）
SQL> select id, a, b, c
2 from (
3 select 'ID' id, max(decode(rn, 1, column_name)) a, max(decode(rn, 2, column_name)) b, max(decode(rn, 3, column_name)) c, 1 s
4 from (
5 select id, extractValue(value(r),'item/column[@name = "ATTRIBUTENAME"]') column_name
6 , extractValue(value(r),'item/column[@name = "MAX(ATTRIBUTEVALUE)"]') value
7 , row_number() over(partition by id order by null) rn
8 from t,
9 table(XMLSequence((t.data).extract('PivotSet/item'))) r
10 )
11 group by id
12 union all
13 select to_char(id), max(decode(rn, 1, value)), max(decode(rn, 2, value)), max(decode(rn, 3, value)) c, 2 s
14 from (
15 select id, extractValue(value(r),'item/column[@name = "ATTRIBUTENAME"]') column_name
16 , extractValue(value(r),'item/column[@name = "MAX(ATTRIBUTEVALUE)"]') value
17 , row_number() over(partition by id order by null) rn
18 from t,
19 table(XMLSequence((t.data).extract('PivotSet/item'))) r
20 )
21 group by id
22 ) order by id, s
23 /
ID A B C
---------------------------------------- -------- -------- --------
ID test test1 test2
1 testVal test1Val test2Val
（由于11.2.3上的ORA-00600，不能使用保理条款）