убрал : из папок

This commit is contained in:
mnoskov
2026-05-28 18:00:56 +03:00
parent e0b74cdc13
commit f045be8a39
30 changed files with 0 additions and 0 deletions
@@ -0,0 +1,45 @@
--DROP TABLE IF EXISTS users;
CREATE TABLE users (
id bigint GENERATED ALWAYS AS IDENTITY PRIMARY KEY,
name varchar(100) NOT NULL,
age SMALLINT,
email varchar(100),
registration_date timestamptz NOT NULL
);
--DROP TABLE IF EXISTS courses;
CREATE TABLE courses (
id bigint GENERATED ALWAYS AS IDENTITY PRIMARY KEY,
title varchar(150) NOT NULL,
category varchar(50),
created_at timestamptz NOT NULL
);
--DROP TABLE IF EXISTS lessons;
CREATE TABLE lessons (
id bigint GENERATED ALWAYS AS IDENTITY PRIMARY KEY,
course_id bigint NOT NULL,
title varchar(150) NOT NULL,
duration_min SMALLINT NOT NULL,
FOREIGN KEY (course_id) REFERENCES courses(id)
);
--DROP TABLE IF EXISTS enrollments;
CREATE TABLE enrollments (
id bigint GENERATED ALWAYS AS IDENTITY PRIMARY KEY,
user_id bigint NOT NULL,
course_id bigint NOT NULL,
enrolled_at timestamptz NOT NULL,
FOREIGN KEY (user_id) REFERENCES users(id),
FOREIGN KEY (course_id) REFERENCES courses(id)
);
--DROP TABLE IF EXISTS lesson_views;
CREATE TABLE lesson_views (
id bigint GENERATED ALWAYS AS IDENTITY PRIMARY KEY,
user_id bigint NOT NULL,
lesson_id bigint NOT NULL,
viewed_at timestamptz NOT NULL,
FOREIGN KEY (user_id) REFERENCES users(id),
FOREIGN KEY (lesson_id) REFERENCES lessons(id)
);
@@ -0,0 +1,155 @@
INSERT INTO users (name, age, email, registration_date) VALUES
('Alice',25,'alice@mail.com','2023-01-10'),
('Bob',30,'bob@gmail.com','2023-02-05'),
('Charlie',22,'charlie@mail.com','2023-02-20'),
('Diana',28,'diana@mail.com','2023-03-01'),
('Ethan',35,'ethan@gmail.com','2023-03-10'),
('Frank',29,'frank@mail.com','2023-01-05'),
('Grace',26,'grace@gmail.com','2023-01-12'),
('Henry',32,'henry@mail.com','2023-01-20'),
('Iris',24,'iris@gmail.com','2023-02-10'),
('Jack',27,'jack@mail.com','2023-04-15'),
('Kelly',23,'kelly@gmail.com','2023-04-20'),
('Leo',31,'leo@mail.com','2023-01-15'),
('Mia',26,'mia@gmail.com','2023-01-18'),
('Nina',28,'nina@mail.com','2023-01-08'),
('Oscar',33,'oscar@gmail.com','2023-01-10'),
('Paul',29,'paul@mail.com','2023-02-15'),
('Quinn',24,'quinn@gmail.com','2023-03-20'),
('Rita',27,'rita@mail.com','2023-04-01'),
('Sam',30,'sam@gmail.com','2023-04-10'),
('Tina',25,'tina@mail.com','2023-04-18');
INSERT INTO courses (title, category, created_at) VALUES
('SQL для начинающих','data','2023-01-01'),
('Python для анализа данных','programming','2023-01-15'),
('BI с нуля','business','2023-03-01'),
('Excel для аналитиков','business','2023-02-01'),
('Основы машинного обучения','data','2023-03-15');
INSERT INTO lessons (course_id, title, duration_min) VALUES
(1,'SELECT и FROM',10),
(1,'JOIN',15),
(1,'WHERE и фильтрация',12),
(1,'Агрегации и GROUP BY',18),
(1,'Подзапросы',25),
(2,'Pandas основы',20),
(2,'NumPy',15),
(2,'Визуализация данных',22),
(3,'Основы BI',12),
(3,'Метрики и дашборды',18),
(3,'Power BI введение',20),
(4,'Формулы в Excel',15),
(4,'Сводные таблицы',18),
(4,'Графики и диаграммы',14),
(5,'Введение в ML',20),
(5,'Линейная регрессия',30),
(5,'Классификация',28),
(5,'Кластеризация',25);
INSERT INTO enrollments (user_id, course_id, enrolled_at) VALUES
(1,1,'2023-01-15'),
(1,2,'2023-02-01'),
(2,1,'2023-01-20'),
(3,2,'2023-03-05'),
(4,3,'2023-04-01'),
(6,1,'2023-01-06'),
(6,2,'2023-01-10'),
(6,4,'2023-02-05'),
(7,1,'2023-01-13'),
(7,3,'2023-03-10'),
(8,1,'2023-01-22'),
(8,2,'2023-01-22'),
(8,3,'2023-01-22'),
(9,4,'2023-02-12'),
(10,1,'2023-04-16'),
(11,2,'2023-04-21'),
(12,1,'2023-01-16'),
(13,2,'2023-01-19'),
(13,4,'2023-02-20'),
(14,1,'2023-01-09'),
(14,2,'2023-01-09'),
(14,3,'2023-01-09'),
(14,4,'2023-02-01'),
(14,5,'2023-03-20'),
(15,1,'2023-01-11'),
(15,3,'2023-03-05'),
(16,2,'2023-02-16'),
(16,4,'2023-02-16'),
(17,3,'2023-03-22'),
(17,5,'2023-03-25'),
(18,1,'2023-04-02'),
(19,1,'2023-04-11'),
(19,2,'2023-04-11'),
(20,2,'2023-04-19'),
(5,5,'2023-03-20'),
(1,4,'2023-02-10'),
(2,3,'2023-03-15'),
(3,1,'2023-02-25'),
(4,2,'2023-03-10'),
(6,3,'2023-03-08'),
(7,2,'2023-02-01'),
(9,1,'2023-02-15'),
(12,3,'2023-03-10'),
(15,2,'2023-02-01'),
(18,4,'2023-04-05');
INSERT INTO lesson_views (user_id, lesson_id, viewed_at) VALUES
(1,1,'2023-01-16 10:00:00'),
(1,2,'2023-01-16 10:15:00'),
(1,3,'2023-01-17 09:00:00'),
(1,4,'2023-01-17 09:30:00'),
(1,4,'2023-01-17 11:00:00'),
(1,4,'2023-01-18 10:00:00'),
(1,6,'2023-02-02 14:00:00'),
(1,12,'2023-02-15 10:00:00'),
(1,13,'2023-02-15 10:30:00'),
(2,1,'2023-01-21 09:00:00'),
(2,9,'2023-03-16 10:00:00'),
(2,10,'2023-03-16 10:30:00'),
(3,6,'2023-03-06 12:00:00'),
(3,1,'2023-02-26 11:00:00'),
(3,2,'2023-02-26 11:20:00'),
(4,9,'2023-04-02 10:00:00'),
(4,10,'2023-04-03 10:00:00'),
(4,6,'2023-03-12 14:00:00'),
(4,7,'2023-03-13 09:00:00'),
(5,15,'2023-03-21 10:00:00'),
(5,16,'2023-03-22 11:00:00'),
(6,1,'2023-01-07 10:00:00'),
(6,2,'2023-01-07 10:15:00'),
(6,3,'2023-01-08 09:00:00'),
(6,4,'2023-01-08 09:30:00'),
(6,5,'2023-01-09 10:00:00'),
(6,6,'2023-01-11 14:00:00'),
(6,7,'2023-01-12 10:00:00'),
(6,8,'2023-01-13 11:00:00'),
(6,12,'2023-02-06 09:00:00'),
(6,13,'2023-02-06 09:30:00'),
(6,14,'2023-02-07 10:00:00'),
(6,9,'2023-03-09 11:00:00'),
(7,1,'2023-01-14 10:00:00'),
(7,2,'2023-01-14 10:20:00'),
(7,3,'2023-01-15 09:00:00'),
(7,9,'2023-03-11 14:00:00'),
(7,10,'2023-03-11 14:30:00'),
(9,12,'2023-02-13 10:00:00'),
(9,13,'2023-02-13 10:30:00'),
(9,1,'2023-02-16 11:00:00'),
(12,1,'2023-01-17 10:00:00'),
(12,2,'2023-01-17 10:20:00'),
(12,1,'2023-04-10 11:00:00'),
(12,3,'2023-04-10 11:30:00'),
(12,9,'2023-03-12 10:00:00'),
(13,6,'2023-01-20 14:00:00'),
(13,7,'2023-01-21 10:00:00'),
(13,12,'2023-02-21 09:00:00'),
(14,1,'2023-01-09 08:00:00'),
(14,2,'2023-01-09 08:15:00'),
(14,3,'2023-01-10 10:00:00'),
(14,4,'2023-01-10 10:30:00'),
(14,5,'2023-01-11 09:00:00'),
(14,6,'2023-01-10 14:00:00'),
(14,7,'2023-01-11 10:00:00'),
(14,8,'2023-01-12 11:00:00'),
(14,9,'2023-01-10 08:00:00');
@@ -0,0 +1,40 @@
CREATE MATERIALIZED VIEW user_cohort_analysis AS
WITH users_and_enrollments_results AS (
SELECT
date_trunc('month', u.registration_date)::date AS cohort_month,
count(DISTINCT u.id) AS users_in_cohort,
count(DISTINCT e.user_id) AS enrolled_users,
count(DISTINCT e.course_id) courses_in_cohort
FROM
users u
JOIN
enrollments e ON u.id = e.user_id
GROUP BY
cohort_month
),
lesson_views_results AS (
SELECT
date_trunc('month', u.registration_date)::date AS cohort_month,
count(DISTINCT lv.user_id) AS active_users,
round(avg(lv.viewed_at::date - u.registration_date::date), 0) AS avg_time_to_first_view
FROM
users u
LEFT JOIN
lesson_views lv ON u.id = lv.user_id
GROUP BY
cohort_month
)
SELECT
uer.cohort_month,
uer.users_in_cohort,
uer.enrolled_users,
lvr.active_users,
round((lvr.active_users / uer.users_in_cohort::numeric) * 100, 2) || '%' AS retention_rate,
round(uer.courses_in_cohort / uer.users_in_cohort::numeric, 2) AS avg_courses_per_user,
lvr.avg_time_to_first_view
FROM
users_and_enrollments_results uer
JOIN
lesson_views_results lvr ON uer.cohort_month = lvr.cohort_month
ORDER BY
uer.cohort_month;
@@ -0,0 +1,114 @@
CREATE MATERIALIZED VIEW conversion_funnel AS
WITH stage_courses AS (
-- registered
(SELECT
0 AS course_id,
1 AS stage_id,
'registered' AS stage_name,
count(DISTINCT u.id) AS users_count
FROM
users u)
UNION ALL
-- enrolled
(SELECT
e.course_id,
2 AS stage_id,
'enrolled' AS stage_name,
count(DISTINCT e.user_id) AS users_count
FROM
enrollments e
GROUP BY
e.course_id
ORDER BY
e.course_id)
UNION ALL
-- started
(SELECT
l.course_id,
3 AS stage_id,
'started' AS stage_name,
count(DISTINCT lv.user_id) AS users_count
FROM
lessons l
JOIN
lesson_views lv ON l.id = lv.lesson_id
GROUP BY
l.course_id
ORDER BY
l.course_id)
UNION ALL
-- mid_progress
(WITH lessons_mid AS (
SELECT
l.course_id,
ceil(percentile_cont(0.5) WITHIN GROUP (ORDER BY l.id)) AS mid_in_lesson_id
FROM
lessons l
GROUP BY
l.course_id
)
SELECT
lm.course_id,
4 AS stage_id,
'mid_progress' AS stage_name,
count(lv.user_id) AS users_count
FROM
lessons_mid lm
LEFT JOIN
lesson_views lv ON lm.mid_in_lesson_id = lv.lesson_id
GROUP BY
lm.course_id, lm.mid_in_lesson_id
ORDER BY
lm.course_id)
UNION ALL
-- completed
(WITH courses_lessons_id AS (
SELECT
l.course_id,
l.id AS lesson_id
FROM
lessons l
GROUP BY
l.course_id, l.id
),
lessons_numbers AS (
SELECT
cli.course_id,
cli.lesson_id,
ROW_NUMBER() OVER(PARTITION BY cli.course_id ORDER BY cli.lesson_id DESC) AS lesson_number
FROM
courses_lessons_id cli
)
SELECT
ln.course_id,
5 AS stage_id,
'completed' AS stage_name,
count(lv.user_id) AS users_count
FROM
lessons_numbers ln
LEFT JOIN
lesson_views lv ON ln.lesson_id = lv.lesson_id
WHERE
ln.lesson_number = 1
GROUP BY
ln.course_id, ln.lesson_id
ORDER BY
ln.course_id)
)
SELECT
sc.course_id,
c.title AS course_title,
sc.stage_name,
sc.users_count,
CASE
WHEN lag(sc.users_count) OVER (PARTITION BY sc.course_id ORDER BY sc.stage_id) IS NULL THEN 'first stage'
WHEN lag(sc.users_count) OVER (PARTITION BY sc.course_id ORDER BY sc.stage_id) = 0 THEN 0::text || '%'
WHEN lag(sc.users_count) OVER (PARTITION BY sc.course_id ORDER BY sc.stage_id) IS NOT NULL THEN round((sc.users_count / lag(sc.users_count) OVER (PARTITION BY sc.course_id ORDER BY sc.stage_id)::NUMERIC) * 100, 0)::text || '%'
END AS conversion_from_previous,
round((sc.users_count / FIRST_VALUE(sc.users_count) OVER (ORDER BY sc.stage_id)::NUMERIC) * 100, 0)::text || '%' AS conversion_from_start
FROM
courses c
RIGHT JOIN
stage_courses sc ON c.id = sc.course_id
ORDER BY
sc.course_id;
@@ -0,0 +1,41 @@
CREATE MATERIALIZED VIEW course_summary AS
WITH courses_unique_users AS (
SELECT
e.course_id,
count(e.user_id) AS unique_users
FROM
enrollments e
GROUP BY
e.course_id
),
lessons_results AS (
SELECT
l.course_id,
count(distinct l.id) AS lessons_count,
count(lv.lesson_id) AS total_views,
min(lv.viewed_at)::date AS first_viewed_at,
max(lv.viewed_at)::date AS last_viewed_at
FROM
lessons l
LEFT JOIN
lesson_views lv ON l.id = lv.lesson_id
GROUP BY
l.course_id
)
SELECT
c.id AS course_id,
c.title AS course_title,
cuu.unique_users,
lr.lessons_count,
lr.total_views,
round(lr.total_views / cuu.unique_users::numeric, 2) AS avg_views_per_user,
lr.first_viewed_at,
lr.last_viewed_at
FROM
courses c
JOIN
courses_unique_users cuu ON c.id = cuu.course_id
JOIN
lessons_results lr ON c.id = lr.course_id
ORDER BY
c.id;
@@ -0,0 +1,39 @@
CREATE MATERIALIZED VIEW lesson_popularity_ranking as
WITH lesson_views_results AS (
SELECT
lv.lesson_id,
count(lv.user_id) AS total_views,
count(DISTINCT lv.user_id) AS unique_viewers
FROM
lesson_views lv
GROUP BY
lv.lesson_id
),
enrollments_results AS (
SELECT
e.course_id,
count(e.user_id) AS enrolled_to_course
FROM
enrollments e
GROUP BY
e.course_id
)
SELECT
l.id AS lesson_id,
l.title AS lesson_title,
c.title AS course_title,
lvr.total_views,
lvr.unique_viewers,
round(lvr.total_views / lvr.unique_viewers::numeric, 2) AS avg_views_per_viewer,
dense_rank() OVER (ORDER BY lvr.total_views desc) AS popularity_rank,
round((lvr.unique_viewers / er.enrolled_to_course::numeric) * 100, 0) || '%' AS completion_rate
FROM
lessons l
JOIN
courses c ON l.course_id = c.id
JOIN
lesson_views_results lvr ON l.id = lvr.lesson_id
JOIN
enrollments_results er ON c.id = er.course_id
ORDER BY
popularity_rank;
@@ -0,0 +1,48 @@
CREATE MATERIALIZED VIEW platform_summary as
WITH users_count AS (
SELECT
count(*) AS total_users
FROM
users
),
courses_count AS (
SELECT
count(*) AS total_courses
FROM
courses
),
lessons_count AS (
SELECT
count(*) AS total_lessons
FROM
lessons
),
users_with_views AS (
SELECT
count(distinct user_id) AS users_with_views
FROM
lesson_views
),
views_count AS (
SELECT
count(*) AS total_views
FROM
lesson_views lv
)
SELECT
uc.total_users,
cc.total_courses,
lc.total_lessons,
uwv.users_with_views,
round(lc.total_lessons / cc.total_courses::NUMERIC, 2) AS avg_lessons_per_course,
round(vc.total_views / lc.total_lessons::NUMERIC, 2) AS avg_views_per_lesson
FROM
users_count uc
CROSS JOIN
courses_count cc
CROSS JOIN
lessons_count lc
CROSS JOIN
users_with_views uwv
CROSS JOIN
views_count AS vc;
@@ -0,0 +1,24 @@
CREATE MATERIALIZED VIEW user_activity_summary AS
SELECT
u.id AS user_id,
u.name,
count(DISTINCT e.course_id) AS enrollment_count,
count(DISTINCT lv.lesson_id) AS viewed_lessons_count,
max(lv.viewed_at)::date AS last_view,
min(lv.viewed_at)::date AS first_view,
((count(DISTINCT lv.lesson_id)::numeric(5,2) / count(DISTINCT l.id)) * 100)::numeric(5, 2) AS engagement_rate
FROM
users u
JOIN
enrollments e ON u.id = e.user_id
JOIN
courses c ON e.course_id = c.id
JOIN
lessons l ON c.id = l.course_id
LEFT JOIN
lesson_views lv ON u.id = lv.user_id -- сюда - 'AND l.id = lv.lesson_id'
GROUP BY
u.id, u.name
ORDER BY
u.id;
-- Кинул его в нейронку на проверку, она сказала что надо добавить 'AND l.id = lv.lesson_id' при присоединении таблицы lesson_views(сделал пометку в коде), это нужно чтобы точно проверить, что урок который посмотрел пользователь точно принадлежит его курсу, чтобы избежать ошибок из-за некорректных данные если такие появятся в результате какого либо сбоя или ошибки.
@@ -0,0 +1,51 @@
-- Задание 1
SELECT
*
FROM
conversion_funnel
WHERE
stage_name = 'started'
ORDER BY
conversion_from_previous;
-- Первые 2 курса, гже отавлилось больше всего учеников между этапами enrolled → started
-- Задание 2
SELECT
*
FROM
lesson_popularity_ranking lpr
WHERE
lpr.avg_views_per_viewer > 1
LIMIT 3;
-- Задание 3
SELECT
*
FROM
user_cohort_analysis uca
ORDER BY
retention_rate DESC;
-- Не разобрался как можжно соеденить с другими таблицам
-- А retention_rate такой, наверное потому что при запуске курсов была пиар компания большая, и делали рассылки с напоминанием позаниматься, а потом перестали, вот и удержание похерилось
-- Задание 4
-- Витрина - user_activity_summary
-- Всяко разно попробовал подобавлять индексы, итог: без индексов быстрее. Просмотрел весь запрос, вроде ничего не нужно менять - лишнего ничего нет.
-- Витрина - course_summary
-- Тут также как и в предидущем - и так и сяк попробовал, план запроса не меняется, время не улучшается
-- Витрина - platform_summary
-- Тут тоже нечего менять - всё из CROSS JOIN. Попробовал вместо count(*) писать count(id) думал мол в ширину будет меньше читать, в итоге выяснил что и так и так одинаково читает
-- Витрина - user_cohort_analysis
-- И тут че не делай, все одна фигня. План вроде хороший.
-- Витрина - lesson_popularity_ranking
-- Тут тоже ничего не вышло - я начниаю чувствовать себя тупым от того что не могу улучшить нигде время запроса.
-- Витрина - conversion_funnel
-- Здесь тоже ничего не вышло - со стороны план запроса выглядит хорошо. Ускорить время запроса не вышло.