several perl-processes updating one table -> collision

Hi people,

I got a tiny little problem with perl, working on an ORA-DB. I start several (up to 8) perl-processes which all work on one table. Each process among them shall be able to read a row from the table, mark it as processed and work down what ever to do. First I tried to do it the following way:

1. select the next item from table
2. update this row to mark it as processed.

Ok... the problem which occurred was, that during selecting and updating from one process, a second process selected the same item and did also his job with that row. So I decided to go another way:

1. update the next item from table with an unique id (pid of perl-process) using the following statement:
   "update test set processed=[pid] where processed=0 and rownum between 0 and 1"
   
2. select all information from this row:
   "select * from test where processed=[pid]"
   
3. and update it to be done:
   "update test set processed=1 where processed=[pid]"
   

Well... the number of collisions was hardly reduced, but in the end there still remained some collisions.

Now my question is: What to do in this case... ? Doesn't Oracle manage those update-statements in a sequential way ?

Hopin' for any ideas...

Regards.
Marcus

-- 
__________________________________________________________
News suchen, lesen, schreiben mit http://newsgroups.web.de