Hello,

I want to execute delete and insert queries thru db2bp and I need to roll-back all the work if an error occured.

Is-it possible to do it with the db2bp ?

Thanks in advance

Technical environment :

DB2 UDB V7.3 FP10
Linux Red Hat 7.3

There are multiple options :

1) Using 'UPDATE COMMAND OPTIONS USING c off".
Go to the db2 prompt and issue the above command, Any command issued following this will not be committed unless you issue an explict commit.

This works only if you are in a db2 prompt.

2) Using export DB2OPTIONS="+c" .

Until you reset the DB2OPTIONS environment variable, your shell w.r. to db2 is autocommit off

3) Using db2 +c
Issuing
db2 +c <SQL STmat>

will mean that this particular SQL STatement is not autocommit.

Use the option that suits you ...

I generally prefer the third one.

Cheers
Sathyaram

what I am getting is that you want to use db2 in command line mode.
db2bp is the back process, every time when you start db2 session, it creates a db2bp for the session(not for each command) and definatly a front process that correspond with that.
you can use

db2 +c insert ...
db2 +c do more work

any time you start db2 without +c ,any commands that where issued with autocommit off will be automatically committed..
Now you can explicitly commit or rollback you commands.

That's I want to do but with an input file.

For exemple :

* input file : test.sql :

connect to test;
create table my_table ( col int );
insert into my_table values ( 1 );
insert into my_table values ( a );
commit;
terminate;

* Command tested :

db2 +a -f -t test.sql

The second insert generates an error ( data type mismatch ) but the first value is inserted is still in my_table.

You may have to use both -s and +c option

db2 +c +a -stf test.sql

Thanks, it works fine now.