Question

I am performing an update on a file that contains over 20 billion records through sql for as400.

   UPDATE "LIBRARY"/AR#RMTPRL9 
   SET SENTFLAG = ' ', DATESENT = '0001-01-01', TIMESENT = '00:00:00', XMITT# = 0, LOCATION = 'PACI175A', ARBATCH# = ' ‘ 
   WHERE LOCATION = 'PACI173A' AND ARBATCH# = 'PAR1981181'

How can I run this update for different values of arbatch# in a single query to save time?

eg: I would like to use the same update statement for ARBATCH# values PAR19811284, PAR18736765 etc.

REEDITS:

I have the following data:(Move these values to location EFG)

  LOCATION     ARBATCH#
  ABC          1234
  ABC          5468
  ABC          8765
  XYZ          6759
  XYZ          5432

Here i would like to achieve a query which would perform update for records as follows:

 UPDATE "LIBRARY"/AR#RMTPRL9 
   SET SENTFLAG = ' ', DATESENT = '0001-01-01', TIMESENT = '00:00:00', XMITT# = 0, LOCATION = 'EFG', ARBATCH# = ' ‘ 
   WHERE LOCATION = 'ABC' AND ARBATCH# = '1234'

Now how could i successfully use “in” for such an update

Answer 1

You can ‘cheat’ pretty trivially:

UPDATE "LIBRARY"/AR#RMTPRL9 
SET SENTFLAG = ' ', 
    DATESENT = '0001-01-01', TIMESENT = '00:00:00', 
    XMITT# = 0, 
    LOCATION = 'EFG', 
    ARBATCH# = ' ' 
WHERE (LOCATION, ARBATCH#) IN (VALUES ('ABC', '1234'),
                                      ('ABC', '5468'),
                                      ('ABC', '8765'),
                                      ('XYZ', '6759'),
                                      ('XYZ', '5432'))

(Statement run on my local iSeries DB2 instance)

Incidently, DATESENT and TIMESENT should be combined into a single timestamp, and probably be nullable, too. I’d personally prefer something like sentAt for the name, not something including the type of the column.