有時候我們需要查詢分區表確定分區的大小。就需要查詢user_tab_partitions 視圖的 HIGH_VALUE字段。
但是會發現這個字段的類型 是LONG 類型。例如：
SQL> desc user_tab_partitions;
Name Type Nullable Default Comments
---------------------- ------------ -------- ------- --------
TABLE_NAME VARCHAR2(30) Y
COMPOSITE VARCHAR2(3) Y
PARTITION_NAME VARCHAR2(30) Y
SUBPARTITION_COUNT NUMBER Y
HIGH_VALUE LONG Y
HIGH_VALUE_LENGTH NUMBER Y
PARTITION_POSITION NUMBER Y
TABLESPACE_NAME VARCHAR2(30) Y
sqlplus 裏面顯示是TO_DATE 具體的內容：
SQL> select t.table_name, t.partition_name, t.high_value
2 from user_tab_partitions t
3 where t.table_name = 'PART_IND_TEST'
4 and t.partition_name = 'P_201401';
TABLE_NAME PARTITION_NAME HIGH_VALUE
------------------------------ ------------------------------ ------------------------------------------------------------------------------------------
PART_IND_TEST P_201401 TO_DATE(' 2014-02-01 00:00:00', 'SYYYY-MM-DD HH24:MI:SS', 'NLS_CALENDAR=GREGORIAN')
查詢得到的內容是 TO_DATE(' 2014-02-01 00:00:00', 'SYYYY-MM-DD HH24:MI:SS', 'NLS_CALENDAR=GREGORIAN')
但是更多的時候需求是希望得到2014-02-01 00:00:00 這個時間。
並且在PL/SQL Developer 查詢出來不是顯示具體的內容
創建一個函數，可以巧妙的繞過這個問題：
create or replace function fun_high_value(p_tname varchar2,
p_parname varchar2)
return varchar2 is
l_high_value varchar2(1000);
begin
select t.high_value
into l_high_value
from user_tab_partitions t
where t.table_name = p_tname
and t.partition_name = p_parname;
return l_high_value;
end;
select table_name,
partition_name,
fun_high_value(table_name, partition_name) high_value,
substr(fun_high_value(table_name, partition_name), 10, 11)
from user_tab_partitions t
where t.table_name = 'PART_IND_TEST'
and t.partition_name = 'P_201401';
TABLE_NAME PARTITION_NAME HIGH_VALUE SUBSTR(FUN_HIGH_VALUE(TABLE_NAME,PARTITION_N
------------------------------ ------------------------------ ------------------------------------------------------------------------------------------ --------------------------------------------
PART_IND_TEST P_201401 TO_DATE(' 2014-02-01 00:00:00', 'SYYYY-MM-DD HH24:MI:SS', 'NLS_CALENDAR=GREGORIAN') 2014-02-01
另外一個辦法是使用XML db的語法查詢。也能實現同樣的需求：
select *
from (with xml as (SELECT dbms_xmlgen.getXMLType('select table_name, partition_name, high_value
from user_tab_partitions t
where t.table_name = ''' ||
t.table_name || ''' ') as x
from user_tables t
where t.TABLE_NAME in ('PART_IND_TEST'))
select extractvalue(rws.object_value, '/ROW/TABLE_NAME') TABLE_NAME,
extractvalue(rws.object_value, '/ROW/PARTITION_NAME') PARTITION_NAME,
extractvalue(rws.object_value, '/ROW/HIGH_VALUE') HIGH_VALUE
FROM XML X, TABLE(XMLSEQUENCE(EXTRACT(X.X, '/ROWSET/ROW'))) rws)
where partition_name = 'P_201401'
如果不想創建函數，數據庫版本是12c以上的話，可以使用with語法：
with function fun_high_value(p_tname varchar2,
p_parname varchar2) return varchar2 is l_high_value varchar2(1000);
begin
select t.high_value
into l_high_value
from user_tab_partitions t
where t.table_name = p_tname
and t.partition_name = p_parname;
return l_high_value;
end;
select table_name,
partition_name,
fun_high_value(table_name, partition_name) high_val
from user_tab_partitions t
where t.table_name = 'PART_IND_TEST'
and t.partition_name = 'P_201401'
以上是幾種辦法。你學會了麼？
下面是我的公衆號：定期會更新一些Oracle 數據庫及Python相關的學習經驗。歡迎關注：
