Обсуждение: A question about the number of times a trigger will fire
I am in the process of trying to set up a function.trigger pair to create
functionally an updateable view. 
I would think that the trigger would fire the number of times that the
calling statement's WHERE clause seceded. Is this incorrect>
I have a view called purchase_view, one of the tables in it's join is a
table called bom_item.
I have defined this trigger:
REATE TRIGGER test_v_trig
    INSTEAD OF INSERT OR UPDATE ON purchase_view
    FOR EACH ROW EXECUTE PROCEDURE v_trig_test();
Prior to firing this trigger this query
select count(*)
FROM purchase_view
WHERE
        proj_no = 3124
        AND
        m_name = 'Mencom' ;
returns 11
The resultant statement generated by the function called by this trigger
is:
UPDATE BOM_ITEM SET  cost_per_unit = 23.45 , qty = 12345.00 
Which in retrospect dies not do what I had in mind :-)
So, it appears that I need to create a WHERE clause for the resultant
statement. But I do not see how the function has enough data to use to
create this where clause.
What am I missing, here?
-- 
"They that would give up essential liberty for temporary safety deserve
neither liberty nor safety."
                        -- Benjamin Franklin
			
		On Thu, Mar 5, 2020 at 7:58 AM stan <stanb@panix.com> wrote:
UPDATE BOM_ITEM SET cost_per_unit = 23.45 , qty = 12345.00
So, it appears that I need to create a WHERE clause for the resultant
statement. But I do not see how the function has enough data to use to
create this where clause.
What am I missing, here?
OLD and/or NEW?
David J.
On Thu, Mar 05, 2020 at 08:58:32AM -0700, David G. Johnston wrote: > On Thu, Mar 5, 2020 at 7:58 AM stan <stanb@panix.com> wrote: > > > UPDATE BOM_ITEM SET cost_per_unit = 23.45 , qty = 12345.00 > > > > So, it appears that I need to create a WHERE clause for the resultant > > statement. But I do not see how the function has enough data to use to > > create this where clause. > > > > What am I missing, here? > > > > > OLD and/or NEW? > > https://www.postgresql.org/docs/12/plpgsql-trigger.html#PLPGSQL-DML-TRIGGER Yes, I was thinking that might be what I had to do. My thinking right now is to create a WHERE clause for the new statement, using (perhaps a subset) of the columns returned by this. -- "They that would give up essential liberty for temporary safety deserve neither liberty nor safety." -- Benjamin Franklin