These are some very nice tips! One of the things I like so much about ActiveRecord is that it plays nice with pretty much every example here, e.g. `Product.wher

These are some very nice tips! One of the things I like so much about ActiveRecord is that it plays nice with pretty much every example here, e.g. `Product.where("title ~* 'title'")` or `Product.where("title ~* ?", title)`. A few other remarks: - I'm so glad to see someone talking about `generate_series` for time-series reports. I often see queries that are `GROUP BY date`, which is usually okay, but will skip an output row if you have no data for a certain day. If you want to force one-row-per-day, use `generate_series`. - His `GROUP BY` query can be rewritten with less repetition like this: SELECT sku, SUM(price), date_part('month', created_at) m FROM invoice_items GROUP BY m HAVING m = 9 In other words, `GROUP BY` and `HAVING` can reference expressions that are aliased in your `SELECT`. - Also on this query: In Postgres if `sku` is a primary key, then you can `SELECT` other columns from `invoice_items` without including them in `GROUP BY`, e.g. `SELECT sku, name, supplier_id, etc`. I think even joined stuff is okay: `SELECT sku, suppliers.name`. - One more nit on this query: I think this is backwards from what he meant?: "having to be sure that whatever you GROUP BY is in your SELECT clause." - Usually I don't use window functions where a `GROUP BY` will do, but to me the biggest value in window functions is they don't force any particular structure on the overall query, so they don't constrain you the way `GROUP BY` does. He kind of alludes to this when he shows per-sku monthly sales combined with total monthly sales. Good example! - I'm not sure his CTE example shows much added value from the CTE, but maybe I'm missing something. The three places I really like CTEs are (1) recursion, (2) with INSERT/UPDATE/DELETE and RETURNING, (3) very big queries. But a lot there depends on taste. - When you use `string_agg` and friends, you can give an ordering for just that function: `string_agg(downloads.title, ',' ORDER BY products.sku)`. That syntax works for every aggregate function (I believe), although it's especially useful for the *_agg family. Thanks for sharing these tips! I wish more people knew the cool things you can do beyond SELECT x FROM y.