oracle scheduler to rename interval partitions - oracle

Question

I get some strange behaviour.
I have a procedure which changes in the names of partitions in a table.
I created a job which runs this procedure every 2 mins for testing.
First run goes very smooth and without any error so far.
However, from 2nd run I get the following error
"ORA-00054: resource busy and acquire with NOWAIT specified or timeout expired
ORA-06512: at "CDS_USER.RENAMING_PARTITIONS", line 17"
Yes, My table is live. But when I run my procedure explicitly withoug the scheduler, it runs perfectly fine. No issues with that. And every first run by the scheduler is also perfect.
Here is the code for scheduler
begin
bms_Scheduler.create_job(
job_name => 'drop_user1' ,
job_action => 'RENAMING_PARTITIONS'
job_action => 'DROP_PARTITIONS'
,start_date => SYSDATE ,
repeat_interval => 'freq=hourly;INTERVAL=7',bysecond=0;' ,
enabled => TRUE ,
comments => 'schduling drop job.');
END;
Here is the code for procedure
create or replace PROCEDURE RENAMING_PARTITIONS
AS
hv varchar2(9);
max_part VARCHAR2(9);
begin
select max(partition_position) into max_part from user_tab_partitions where table_name='DEMO';
for x in (select partition_name, high_value, partition_position
from user_tab_partitions
where table_name = 'DEMO' and partition_name like 'SYS%')
loop
if x.partition_position <> max_part THEN
execute immediate 'select to_char('||x.high_value||'-1,''YYYYMMDD'') from dual' into hv;
partition '||x.partition_name
--||' to DATE_'||hv);
execute immediate('alter table DEMO rename partition '||x.partition_name
||' to DATE_'||hv);
end if;
end loop;
end;
How can I fix this?

Answer 1

Don't do this. I know its inconvenient that Oracle won't follow a predefined partition naming schema for system created interval partitions (probably an enhancement to come in future), but altering the table every 2 min is not a good alternative by any stretch.
Luckily, you really only need to know the partition name in advance when specifying it in queries. The typical query would be something like:
select * from my_table partition (blah);
In 11g, you can use the "partition for" clause to workaround the system generated names, as follows (example for interval DATE partitions):
SQL> set display on
SQL> set linesize 200
SQL> drop table test_data
Table dropped.
SQL> create table test_data (
start_date DATE,
store_id NUMBER,
inventory_id NUMBER,
qty_sold NUMBER
)
PARTITION BY RANGE (start_date)
INTERVAL(NUMTOYMINTERVAL(1, 'MONTH'))
(
PARTITION part_01 values LESS THAN (TO_DATE('20130101','YYYYMMDD'))
)
Table created.
SQL> insert into test_data values (to_date('20121231', 'YYYYMMDD'), 1, 2, 1)
1 row created.
SQL> commit
Commit complete.
SQL> SELECT table_name, partition_name, high_value
FROM user_tab_partitions
WHERE table_name = 'TEST_DATA'
ORDER BY table_name, partition_name
TABLE_NAME PARTITION_NAME HIGH_VALUE
------------------------------ ------------------------------ --------------------------------------------------
TEST_DATA PART_01 TO_DATE(' 2013-01-01 00:00:00', 'SYYYY-MM-DD HH24:
MI:SS', 'NLS_CALENDAR=GREGORIAN')
1 row selected.
SQL> insert into test_data values (to_date('20130101', 'YYYYMMDD'), 1, 5, 8)
1 row created.
SQL> insert into test_data values (to_date('20130115', 'YYYYMMDD'), 2, 4, 5)
1 row created.
SQL> insert into test_data values (sysdate, 2, 3, 2)
1 row created.
SQL> commit
Commit complete.
SQL> SELECT table_name, partition_name, high_value
FROM user_tab_partitions
WHERE table_name = 'TEST_DATA'
ORDER BY table_name, partition_name
TABLE_NAME PARTITION_NAME HIGH_VALUE
------------------------------ ------------------------------ --------------------------------------------------
TEST_DATA PART_01 TO_DATE(' 2013-01-01 00:00:00', 'SYYYY-MM-DD HH24:
MI:SS', 'NLS_CALENDAR=GREGORIAN')
TEST_DATA SYS_P67 TO_DATE(' 2013-02-01 00:00:00', 'SYYYY-MM-DD HH24:
MI:SS', 'NLS_CALENDAR=GREGORIAN')
TEST_DATA SYS_P68 TO_DATE(' 2013-03-01 00:00:00', 'SYYYY-MM-DD HH24:
MI:SS', 'NLS_CALENDAR=GREGORIAN')
3 rows selected.
SQL> -- get data for January partition only
SQL> select * from test_data partition for (to_date('20130101', 'YYYYMMDD'))
START_DATE STORE_ID INVENTORY_ID QTY_SOLD
----------- ---------- ------------ ----------
01-JAN-2013 1 5 8
15-JAN-2013 2 4 5
2 rows selected.