In a previous post, we have provided an example of how to generate date series in PostgreSQL. Let’s see how we can do it in Redshift. Here, it is not so straightforward since we have to generate a series first. A good example is to consider an arbitrary table to get the row numbers and then based on the row numbers to generate the date series. Let’s say that we want to generate a monthly date series from now and looking behind 200 months.
SELECT
DATEADD('month', -n, DATE_TRUNC('month', add_months(CURRENT_DATE,12))) AS gdate
FROM (SELECT ROW_NUMBER() OVER () AS n FROM one_table LIMIT 200) n
Notice: So far you cannot use the generate_series() in Redshift because it is supported only by the master node and it fails to work when you have to join it with other tables
