# you have table1 and table 2 with same schema
select ref, "deleted" as diff from table1
    where ref not in (select ref from table2)
union
select ref, "new" as diff from table2
    where ref not in (select ref from table1)
union
# if more than one column is needed to check if record changed, then you can add it in the `case` statement
# in this example we only have "val" column, but let's say you have "val", "val2" and "val3" columns
# then, to check if record changed the case statement should look like this
# (case when t1.val<>t2.val or t1.val2<>t2.val or t1.val3<>t2.val<>3 then 'changed' else 'same' end)
# then, we would end with a `select` statement looking like this:
# select t1.ref, (case when t1.val<>t2.val or t1.val2<>t2.val or t1.val3<>t2.val<>3 then 'changed' else 'same' end) as diff from table1 t1
select t1.ref, (case when t1.val<>t2.val then 'changed' else 'same' end) as diff from table1 t1
    left join table2 t2 on t2.ref=t1.ref
    where t2.ref=t1.ref;