I'd like to update six rows of table "scores" at once using a "with"
and a "(values" construct as shown below.

Actually I'm open to any construct that does the job.

This statement returns a "SQL0104N An unexpected token "a" was found
following scores" error message.

Can it be tweaked to work?

Thanks.

with b(p_id,r_id,z_id,score_no)
as(values (201101954,3,1,0.20),
(201101954,3,4,0.20),
(201101954,3,5,0.10),
(201101954,3,6,0.05),
(201101954,3,7,0.07),
(201101954,3,8,0.09))
update scores a,b
set a.score_no=b.score_no
where
a.p_id=b.p_id
and a.r_id=b.r_id
and a.z_id=b.z_id

Your update statement violated syntax of UPDATE statenet in DB2.

Please try...

UPDATE scores a
SET score_no
= (SELECT b.score_no
FROM (VALUES (201101954 , 3 , 1 , 0.20)
, (201101954 , 3 , 4 , 0.20)
, (201101954 , 3 , 5 , 0.10)
, (201101954 , 3 , 6 , 0.05)
, (201101954 , 3 , 7 , 0.07)
, (201101954 , 3 , 8 , 0.09)
) b(p_id , r_id , z_id , score_no)
WHERE a.p_id = b.p_id
AND a.r_id = b.r_id
AND a.z_id = b.z_id
)
;

score_no would be set to null in my last example, if the row in a
doesn't match with b.

To keep values of score_no at the time the row doesn't match with b.
UPDATE scores a
SET score_no
= COALESCE(
(SELECT b.score_no
FROM (VALUES (201101954 , 3 , 1 , 0.20)
, (201101954 , 3 , 4 , 0.20)
, (201101954 , 3 , 5 , 0.10)
, (201101954 , 3 , 6 , 0.05)
, (201101954 , 3 , 7 , 0.07)
, (201101954 , 3 , 8 , 0.09)
) b(p_id , r_id , z_id , score_no)
WHERE a.p_id = b.p_id
AND a.r_id = b.r_id
AND a.z_id = b.z_id
)
, a.score_no
)
;

Or, it would be better to use a MERGE statement.

MERGE scores a
USING (VALUES (201101954 , 3 , 1 , 0.20)
, (201101954 , 3 , 4 , 0.20)
, (201101954 , 3 , 5 , 0.10)
, (201101954 , 3 , 6 , 0.05)
, (201101954 , 3 , 7 , 0.07)
, (201101954 , 3 , 8 , 0.09)
) b(p_id , r_id , z_id , score_no)
ON b.p_id = a.p_id
AND b.r_id = a.r_id
AND b.z_id = a.z_id
WHEN MATCHED THEN
UPDATE
SET score_no = b.score_no
;