How to sort by timestamp streaks SQL - mysql

Question

Say I have the following tables (simplified version of what I'm working with):
CREATE TABLE posts (
id INTEGER,
title VARCHAR(255),
text TEXT,
author_id INTEGER,
created_at TIMESTAMP
);
CREATE TABLE authors (
id INTEGER,
name VARCHAR(255),
email VARCHAR(255)
);
What I want to do is retrieve only the authors but order them by number of one-week streaks. That is, number of consecutive weeks an author published a post. The time a post has been made is stored in the posts created_at column
What I'm having the most difficulty with is understanding how to calculate the difference in time between posts across rows. I’m using MySQL

Answer 1

Without window functions it's a bit hard to do in MySql 5.7
But here's an experiment test snippet that uses variables :
Sample data:
DROP TABLE IF EXISTS `posts`;
DROP TABLE IF EXISTS `authors`;
CREATE TABLE `authors` (
id INTEGER PRIMARY KEY AUTO_INCREMENT,
name VARCHAR(255),
email VARCHAR(255)
);
CREATE TABLE `posts` (
id INTEGER PRIMARY KEY AUTO_INCREMENT,
title VARCHAR(255),
`text` TEXT,
author_id INTEGER,
created_at TIMESTAMP,
CONSTRAINT fk_posts_author_id FOREIGN KEY (author_id) REFERENCES `authors`(id)
);
insert into `authors` (name, email) values
('john doe', 'john.doe#home.net'),
('jane sheppard', 'jane.sheppard#home.net');
insert into `posts` (author_id, created_at, title, `text`) values
(1, '2019-02-07', 'When', 'bla'),
(1, '2019-02-09', 'I', 'bla2'),
(1, '2019-02-14', 'Start', 'bla3'),
(1, '2019-02-19', 'looking', 'bla4'),
(1, '2019-03-10', '...', 'bla5'),
(2, '2019-02-01', 'I', 'blah1'),
(2, '2019-02-05', 'frighten', 'blah2'),
(2, '2019-02-19', 'even', 'blah3'),
(2, '2019-03-20', 'myself', 'blah4');
Query:
SELECT q3.ConcurrentWeeks, q3.StartWeekDate, a.*
FROM
(
SELECT COUNT(*) as ConcurrentWeeks, MIN(WkDt) as StartWeekDate, author_id
FROM
(
SELECT q1.WkDt, q1.Total
, case
when #author = author_id and #yr = yr and #wk = wk-1 then #rnk
else #rnk := #rnk + 1
end as rnk
, #author := author_id as author_id
, #yr := yr as yr
, #wk := wk as wk
FROM
(
SELECT
author_id, YEAR(created_at) as yr, WEEK(created_at) as wk
, COUNT(*) AS Total
, COALESCE(MIN(STR_TO_DATE(concat(YEAR(created_at),' monday ',WEEK(created_at)),'%X %W %V')), MIN(CAST(created_at AS DATE))) AS WkDt
FROM `posts` p
GROUP BY author_id, YEAR(created_at), WEEK(created_at)
ORDER BY author_id, yr, wk
) q1
CROSS JOIN (select #author := null, #yr := null, #wk := null, #rnk := 0) init
) q2
GROUP BY author_id, rnk
HAVING ConcurrentWeeks > 1
) q3
LEFT JOIN `authors` a ON a.id = q3.author_id
ORDER BY ConcurrentWeeks DESC, StartWeekDate ASC
Result:
ConcurrentWeeks StartWeekDate id name email
--------------- ------------- -- ------------- ----------------------
3 2019-02-04 1 john doe john.doe#home.net
2 2019-01-28 2 jane sheppard jane.sheppard#home.net