That could be possible. An easier approach is to modify the queries to return a row for each category. That can be done by changing the final GROUP BY clause to include the category ID. I haven’t tried this with all the examples that I posted, but here’s a modification of the “Post counts for interval” query that does that:
--[params]
-- string :query_interval = 7 days
-- date :start_date
-- date :end_date
-- int_list :category_ids = -1
-- boolean :include_subcategories = true
WITH query_periods AS (
SELECT generate_series(:start_date, :end_date, :query_interval::interval)::date as period_start
),
subcategory_ids AS (
SELECT id FROM categories
WHERE parent_category_id IN (:category_ids)
),
sub_subcategory_ids AS (
SELECT id FROM categories
WHERE parent_category_id IN (SELECT id FROM subcategory_ids)
)
SELECT
period_start,
t.category_id,
COUNT(p.id)
FROM query_periods qp
JOIN posts p
ON p.created_at::date >= qp.period_start AND p.created_at::date < qp.period_start + interval :query_interval
JOIN topics t
ON t.id = p.topic_id
WHERE t.archetype = 'regular'
AND t.deleted_at IS NULL
AND CASE
WHEN -1 IN (:category_ids)
THEN true
WHEN :include_subcategories = false
THEN t.category_id IN (:category_ids)
ELSE t.category_id IN (:category_ids) OR t.category_id IN (SELECT id FROM subcategory_ids) OR t.category_id IN (SELECT id FROM sub_subcategory_ids)
END
AND p.deleted_at IS NULL
AND p.post_type = 1
GROUP BY period_start, t.category_id
ORDER BY period_start
Here’s what the results look like on my development site:
