Updating an is_latest column on redshift table - sql

Question

I'm attempting to update an is_latest column on a redshift table grouped by source and source_primary_key with the following update statement, but getting an error that window functions are not allowed in update statements. What is the best way to go about this?
update my_schema.production_log
set is_latest =
case when run_start_time = max(run_start_time) over (partition by
source, source_primary_key)
then ‘t’ else ‘f’
end

Answer 1

Redshift supports a FROM clause in the update statement. Try this:
update my_schema.production_log
set is_latest = (case when run_start_time = max_run_start_time
then 't' else 'f'
end)
from (select source, source_primary_key, max(run_start_time) as max_run_start_time
from my_schema.production_log
group by source, source_primary_key
) pl2
where pl2.source = production_log.source and
pl2.source_primary_key = production_log.source_primary_key;

Answer 2

You need to rewrite the query so it can join to a derived table:
UPDATE u
SET u.is_latest =CASE
WHEN u.run_start_time = j.max_time THEN 't'
ELSE 'f'
END
FROM my_schema.production_log AS u
INNER JOIN (
SELECT
source
, source_primary_key
, MAX( run_start_time ) max_time
FROM my_schema.production_log
GROUP BY
source
, source_primary_key
) AS j
ON u.source = j.source
AND u.source_primary_key = j.source_primary_key