I have a script that uses a stack of with clauses to come up with some result, and then I want to write that result in a table. I just can\'t get my head around it, could so
You can always do something like this:
update mytable t
set SomeColumn = c.ComputedValue
from (select *, 42 as ComputedValue from mytable where id = 1) c
where t.id = c.id
You can now also use with statement inside update
update mytable t
set SomeColumn = c.ComputedValue
from (with abc as (select *, 43 as ComputedValue_new from mytable where id = 1
select *, 42 as ComputedValue, abc.ComputedValue_new from mytable n1
inner join abc on n1.id=abc.id) c
where t.id = c.id
The WITH syntax appears to be valid in an inline view, e.g.
UPDATE (WITH comp AS ...
SELECT SomeColumn, ComputedValue FROM t INNER JOIN comp ...)
SET SomeColumn=ComputedValue;
But in the quick tests I did this always failed with ORA-01732: data manipulation operation not legal on this view
, although it succeeded if I rewrote to eliminate the WITH clause. So the refactoring may interfere with Oracle's ability to guarantee key-preservation.
You should be able to use a MERGE, though. Using the simple example you've posted this doesn't even require a WITH clause:
MERGE INTO mytable t
USING (select *, 42 as ComputedValue from mytable where id = 1) comp
ON (t.id = comp.id)
WHEN MATCHED THEN UPDATE SET SomeColumn=ComputedValue;
But I understand you have a more complex subquery you want to factor out. I think that you will be able to make the subquery in the USING
clause arbitrarily complex, incorporating multiple WITH
clauses.
If anyone comes here after me, this is the answer that worked for me.
NOTE: please make to read the comments before using this, this not complete. The best advice for update queries I can give is to switch to SqlServer ;)
update mytable t
set z = (
with comp as (
select b.*, 42 as computed
from mytable t
where bs_id = 1
)
select c.computed
from comp c
where c.id = t.id
)
Good luck,
GJ