Merge pull request 'add Step 1 from pet project Clickhouse' (#19) from OS-68 into main

Reviewed-on: https://gitea.de-infra.servehttp.com/de-infrastructure/DWH/pulls/19
This commit is contained in:
2025-12-13 13:02:53 +04:00
7 changed files with 418 additions and 0 deletions
+418
View File
@@ -0,0 +1,418 @@
-- Этап 1
-- Задание 1
-- Перенести данные как в пет-проекте с Greenplum
CREATE TABLE itikhomirov.core_user
ON CLUSTER clickhouse_cluster
(
user_id Int32,
name String,
age Int32,
email String,
registration_date DateTime64(3, 'UTC')
)
ENGINE = ReplicatedMergeTree(
'/clickhouse/tables/{shard}/itikhomirov/core_user',
'{replica}'
)
ORDER BY user_id;
ALTER TABLE itikhomirov.core_user
MODIFY COLUMN registration_date DateTime64(3);
CREATE TABLE itikhomirov.core_lesson
ON CLUSTER clickhouse_cluster
(
lesson_id Int32,
title String,
duration_min Int32,
course_id Int32
)
ENGINE = ReplicatedMergeTree(
'/clickhouse/tables/{shard}/itikhomirov/core_lesson',
'{replica}'
)
ORDER BY lesson_id;
CREATE TABLE itikhomirov.core_fact_lesson_views
ON CLUSTER clickhouse_cluster
(
user_id Int32,
lesson_id Int32,
course_id Int32,
viewed_at DateTime64(3)
)
ENGINE = ReplicatedMergeTree(
'/clickhouse/tables/{shard}/itikhomirov/core_fact_lesson_views',
'{replica}'
)
ORDER BY user_id;
ALTER TABLE itikhomirov.core_fact_lesson_views
MODIFY COLUMN viewed_at DateTime64(3);
DROP TABLE IF EXISTS itikhomirov.core_fact_lesson_views
ON CLUSTER clickhouse_cluster;
CREATE TABLE itikhomirov.core_enrollments
ON CLUSTER clickhouse_cluster
(
user_id Int32,
course_id Int32,
enrolled_at DateTime64(3)
)
ENGINE = ReplicatedMergeTree(
'/clickhouse/tables/{shard}/itikhomirov/core_enrollmentsы',
'{replica}'
)
ORDER BY user_id;
CREATE TABLE itikhomirov.core_course
ON CLUSTER clickhouse_cluster
(
course_id Int32,
title String,
category String
)
ENGINE = ReplicatedMergeTree(
'/clickhouse/tables/{shard}/itikhomirov/core_course',
'{replica}'
)
ORDER BY course_id;
-- Этап 2. Создание витринных таблиц
-- Задание 2: Подготовь структуру витрины lesson_popularity_summary
-- Создаем пустую таблицу lesson_popularity_summary(структуру витрины)
CREATE TABLE itikhomirov.lesson_popularity_summary
ON CLUSTER clickhouse_cluster
(
lesson_id UInt32,
lesson_title String,
course_id UInt32,
course_title String,
total_views UInt32,
unique_users UInt32,
first_view DateTime64(3),
last_view DateTime64(3)
)
ENGINE = ReplicatedMergeTree(
'/clickhouse/tables/{shard}/itikhomirov/lesson_popularity_summary',
'{replica}'
)
ORDER BY (lesson_id, course_id);
-- Создаем Materialized View
DROP TABLE IF EXISTS itikhomirov.mv_lesson_popularity_summary
ON CLUSTER clickhouse_cluster;
CREATE MATERIALIZED VIEW itikhomirov.mv_lesson_popularity_summary
ON CLUSTER clickhouse_cluster
TO itikhomirov.lesson_popularity_summary
AS
SELECT
flv.lesson_id AS lesson_id,
l.title AS lesson_title,
flv.course_id AS course_id,
c.title AS course_title,
count() AS total_views,
uniqExact(flv.user_id) AS unique_users,
min(flv.viewed_at) AS first_view,
max(flv.viewed_at) AS last_view
FROM itikhomirov.core_fact_lesson_views AS flv
JOIN itikhomirov.core_lesson AS l ON flv.lesson_id = l.lesson_id
JOIN itikhomirov.core_course AS c ON flv.course_id = c.course_id
GROUP BY
flv.lesson_id,
l.title,
flv.course_id,
c.title;
-- Вставляем данные в таблицу lesson_popularity_summary
INSERT INTO itikhomirov.lesson_popularity_summary
SELECT
flv.lesson_id,
l.title AS lesson_title,
flv.course_id,
c.title AS course_title,
count() AS total_views,
uniqExact(flv.user_id) AS unique_users,
min(flv.viewed_at) AS first_view,
max(flv.viewed_at) AS last_view
FROM itikhomirov.core_fact_lesson_views AS flv
JOIN itikhomirov.core_lesson AS l ON flv.lesson_id = l.lesson_id
JOIN itikhomirov.core_course AS c ON flv.course_id = c.course_id
GROUP BY
flv.lesson_id,
l.title,
flv.course_id,
c.title;
-- Задание 3: Создай витрину inactive_users_summary
-- Создаем пустую таблицу inactive_users_summary(структуру витрины)
DROP TABLE IF EXISTS itikhomirov.inactive_users_summary
ON CLUSTER clickhouse_cluster;
DROP TABLE IF EXISTS itikhomirov.inactive_users_summary
ON CLUSTER clickhouse_cluster;
CREATE TABLE itikhomirov.inactive_users_summary
ON CLUSTER clickhouse_cluster
(
user_id UInt32,
name String,
email String,
age UInt8,
registered_courses_count UInt32,
registration_date DateTime64(3)
)
ENGINE = ReplicatedMergeTree(
'/clickhouse/tables/{shard}/itikhomirov/inactive_users_summary',
'{replica}'
)
ORDER BY (user_id);
-- Создаем Materialized View
DROP TABLE IF EXISTS itikhomirov.mv_inactive_users_summary
ON CLUSTER clickhouse_cluster;
CREATE MATERIALIZED VIEW itikhomirov.mv_inactive_users_summary
ON CLUSTER clickhouse_cluster
TO itikhomirov.inactive_users_summary
AS
WITH user_courses AS (
SELECT
e.user_id,
countDistinct(e.course_id) AS registered_courses_count
FROM itikhomirov.core_enrollments AS e
GROUP BY
e.user_id
)
SELECT
u.user_id AS user_id,
u.name AS name,
u.email AS email,
u.age AS age,
coalesce(uc.registered_courses_count, 0) AS registered_courses_count,
u.registration_date AS registration_date
FROM itikhomirov.core_user AS u
LEFT JOIN user_courses AS uc
ON u.user_id = uc.user_id
WHERE u.user_id NOT IN (
SELECT DISTINCT user_id
FROM itikhomirov.core_fact_lesson_views
);
INSERT INTO itikhomirov.inactive_users_summary
WITH user_courses AS (
SELECT
e.user_id,
countDistinct(e.course_id) AS registered_courses_count
FROM itikhomirov.core_enrollments AS e
GROUP BY
e.user_id
)
-- В CTE user_courses считается число курсов по каждому пользователю независимо от просмотров
-- Снаружи берутся пользователи без просмотров (NOT IN по core_fact_lesson_views),
-- и к ним подмешивается уже готовый registered_courses_count.
SELECT
u.user_id AS user_id,
u.name AS name,
u.email AS email,
u.age AS age,
coalesce(uc.registered_courses_count, 0) AS registered_courses_count,
u.registration_date AS registration_date
FROM itikhomirov.core_user AS u
LEFT JOIN user_courses AS uc
ON u.user_id = uc.user_id
WHERE u.user_id NOT IN (
SELECT DISTINCT user_id
FROM itikhomirov.core_fact_lesson_views
)
ORDER BY u.user_id;
-- Задание 4: Создай витрину course_completion_rate
CREATE TABLE itikhomirov.course_completion_rate
ON CLUSTER clickhouse_cluster
(
user_id UInt32,
user_name String,
course_id UInt32,
course_title String,
lessons_in_course UInt32,
lessons_viewed UInt32,
completion_rate Float32
)
ENGINE = ReplicatedMergeTree(
'/clickhouse/tables/{shard}/itikhomirov/course_completion_rate',
'{replica}'
)
ORDER BY (user_id, course_id);
-- Создаем Materialized View
CREATE MATERIALIZED VIEW itikhomirov.mv_course_completion_rate
ON CLUSTER clickhouse_cluster
TO itikhomirov.course_completion_rate
AS
WITH
lessons_per_course AS (
SELECT
course_id,
countDistinct(lesson_id) AS lessons_in_course
FROM itikhomirov.core_lesson
GROUP BY course_id
),
views_per_user_course AS (
SELECT
user_id,
course_id,
countDistinct(lesson_id) AS lessons_viewed
FROM itikhomirov.core_fact_lesson_views
GROUP BY user_id, course_id
)
SELECT
u.user_id AS user_id,
u.name AS user_name,
c.course_id AS course_id,
c.title AS course_title,
lpc.lessons_in_course AS lessons_in_course,
coalesce(vpc.lessons_viewed, 0) AS lessons_viewed,
ROUND(
COALESCE(
NULLIF(coalesce(vpc.lessons_viewed, 0) * 1.0, 0)
/ NULLIF(lpc.lessons_in_course * 1.0, 0),
0
),
2
) AS completion_rate
FROM itikhomirov.core_user AS u
JOIN itikhomirov.core_enrollments AS e ON u.user_id = e.user_id
JOIN itikhomirov.core_course AS c ON e.course_id = c.course_id
JOIN lessons_per_course AS lpc ON c.course_id = lpc.course_id
LEFT JOIN views_per_user_course AS vpc ON vpc.user_id = u.user_id AND vpc.course_id = c.course_id
ORDER BY
u.user_id,
c.course_id;
-- Вставляем данные в таблицу pet.course_completion_rate
INSERT INTO itikhomirov.course_completion_rate
-- Создаем две временные таблицы:
-- 1) Сколько уроков в каждом курсе
WITH
lessons_per_course AS (
SELECT
course_id,
countDistinct(lesson_id) AS lessons_in_course
FROM itikhomirov.core_lesson
GROUP BY course_id
),
-- 2) Сколько уроков каждого курса посмотрел пользователь
views_per_user_course AS (
SELECT
user_id,
course_id,
countDistinct(lesson_id) AS lessons_viewed
FROM itikhomirov.core_fact_lesson_views
GROUP BY user_id, course_id
)
-- Основной запрос
SELECT
u.user_id AS user_id,
u.name AS user_name,
c.course_id AS course_id,
c.title AS course_title,
lpc.lessons_in_course AS lessons_in_course,
coalesce(vpc.lessons_viewed, 0) AS lessons_viewed,
ROUND(
COALESCE(
NULLIF(coalesce(vpc.lessons_viewed, 0) * 1.0, 0)
/ NULLIF(lpc.lessons_in_course * 1.0, 0),
0
),
2
) AS completion_rate
FROM itikhomirov.core_user AS u
JOIN itikhomirov.core_enrollments AS e ON u.user_id = e.user_id
JOIN itikhomirov.core_course AS c ON e.course_id = c.course_id
JOIN lessons_per_course AS lpc ON c.course_id = lpc.course_id
LEFT JOIN views_per_user_course AS vpc ON vpc.user_id = u.user_id AND vpc.course_id = c.course_id
ORDER BY
u.user_id,
c.course_id;
-- Этап 3. Анализ и запросы
-- Задание 5: Напиши SQL-запросы к витринам
-- Уроки с хотя бы одним просмотром
SELECT
lesson_title,
course_title,
total_views
FROM itikhomirov.lesson_popularity_summary
WHERE total_views > 0
ORDER BY lesson_title ASC;
-- Топ‑3 урока внутри одного курса
SELECT
lesson_title,
course_title,
total_views
FROM itikhomirov.lesson_popularity_summary
WHERE course_title = 'SQL для начинающих'
ORDER BY total_views DESC
LIMIT 3;
-- Пользователи, которые завершили курсы на 100%
SELECT
user_id,
user_name,
course_title,
completion_rate
FROM itikhomirov.course_completion_rate
WHERE completion_rate = 1
ORDER BY user_id;
-- Этап 4. Тестирование и валидация
-- Задание 6: Сравни данные с Greenplum**
-- 1. Выбери 2–3 значения из таблиц `_summary_source` в Greenplum
-- 2. Сравни с ClickHouse-результатами:
-- - Совпадает ли `SUM`, `AVG`, `MIN`, `MAX`
-- - Проверяй `NULL`, `0`, округления
-- 3. Напиши мини-отчёт: **"где совпадает, где нет, почему"**
SELECT
sum(total_views) AS sum_total_views,
avg(total_views) AS avg_total_views,
min(total_views) AS min_total_views,
max(total_views) AS max_total_views
FROM itikhomirov.lesson_popularity_summary
WHERE lesson_id IN (1, 3, 5);
SELECT
sum(total_views) AS sum_total_views,
avg(total_views) AS avg_total_views,
min(total_views) AS min_total_views,
max(total_views) AS max_total_views
FROM itikhomirov.lesson_popularity_summary
WHERE lesson_id = 999;
-- Отчёт в .doc файле
Binary file not shown.

After

Width:  |  Height:  |  Size: 292 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 359 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 215 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 129 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 234 KiB

Binary file not shown.