Практика 'Синтаксис' - Clickhouse

2026-05-17 22:14:14 +03:00
parent 21aa48c237
commit 392f412bbd
4 changed files with 215 additions and 0 deletions
@@ -0,0 +1,215 @@
 -- Задание 1
 CREATE TABLE mnoskov.users
 ON CLUSTER clickhouse_cluster
 (
 	user_id UInt32,
 	name String,
 	age UInt8,
 	country String,
 	is_active UInt8
 )
 ENGINE = ReplicatedMergeTree(
    '/clickhouse/tables/{shard}/mnoskov/users',
    '{replica}'
 )
 ORDER BY user_id;
 INSERT INTO mnoskov.users
 (user_id,name,age,country,is_active)
 VALUES
 (1,'Ivan',30,'RU',1),
 (2,'Anna',25,'RU',1),
 (3,'Petr',45,'RU',0),
 (4,'Olga',35,'RU',1),
 (5,'John',55,'US',1),
 (6,'Alice',42,'US',0),
 (7,'Tom',38,'US',1),
 (8,'Lucas',28,'BR',1),
 (9,'Maria',60,'BR',1),
 (10,'Joao',27,'BR',0),
 (11,'Kenji',33,'JP',1),
 (12,'Yuki',29,'JP',0),
 (13,'Chen',40,'CN',1),
 (14,'Li',48,'CN',1),
 (15,'Fatima',31,'AE',1),
 (16,'Ahmed',36,'AE',1),
 (17,'George',24,'US',1),
 (18,'Nina',34,'RU',1),
 (19,'Max',50,'DE',1),
 (20,'Sophie',22,'DE',0);
 SELECT * FROM mnoskov.users;
 --1
 SELECT
 	user_id,
 	age,
 	is_active
 FROM
 	mnoskov.users
 WHERE
 	is_active = TRUE
 ORDER BY
 	age DESC
 LIMIT 10;
 --2
 SELECT
 	name,
 	country
 FROM
 	mnoskov.users
 WHERE
 	country = 'RU'
 ORDER BY
 	name;
 --3
 SELECT
 	user_id,
 	age,
 	is_active
 FROM
 	mnoskov.users
 WHERE
 	is_active = TRUE
 LIMIT 10;
 -- Данные вывелить так как лежат на диске.
 --4
 SELECT
 	country,
 	count()
 FROM
 	mnoskov.users
 GROUP BY
 	country;
 --5
 SELECT
 	country,
 	min(age) AS min_age,
 	max(age) AS max_age
 FROM
 	mnoskov.users
 GROUP BY
 	country;
 --6
 SELECT
 	country,
 	count() AS cnt_users
 FROM
 	mnoskov.users
 GROUP BY
 	country
 ORDER BY
 	cnt_users DESC;
 --Доп
 SELECT
 	uniqExact(country),
 	uniq(country)
 FROM
 	mnoskov.users;
 -- uniqExact(country) - кол-во уникальных значений в колонке country. Но uniq работает также.
 -- В инете написано что uniq быстрее.
 -- Задание 2
 CREATE TABLE mnoskov.events1
 ON CLUSTER clickhouse_cluster
 (
 	user_id UInt32,
 	event_time DateTime,
 	event_type String
 )
 ENGINE = ReplicatedMergeTree(
    '/clickhouse/tables/{shard}/mnoskov/events1',
    '{replica}'
 )
 ORDER BY(user_id,event_time);
 INSERT INTO mnoskov.events1
 (user_id,event_time,event_type)
 VALUES
 (1,'2025-07-17 09:00:00','click'),
 (1,'2025-07-17 09:05:00','view'),
 (1,'2025-07-17 09:10:00','purchase'),
 (2,'2025-07-17 10:00:00','click'),
 (2,'2025-07-17 10:05:00','click'),
 (2,'2025-07-17 11:00:00','view'),
 (3,'2025-07-16 12:00:00','view'),
 (3,'2025-07-16 12:30:00','click'),
 (3,'2025-07-17 08:00:00','purchase'),
 (4,'2025-07-17 13:00:00','click'),
 (4,'2025-07-17 14:00:00','view'),
 (4,'2025-07-17 15:00:00','view'),
 (5,'2025-07-16 20:00:00','view'),
 (5,'2025-07-17 01:00:00','click'),
 (5,'2025-07-17 06:00:00','purchase'),
 (6,'2025-07-17 07:30:00','view'),
 (6,'2025-07-17 08:45:00','click'),
 (7,'2025-07-16 23:59:59','click'),
 (7,'2025-07-17 00:01:00','purchase'),
 (7,'2025-07-17 10:10:00','view');
 SELECT * FROM mnoskov.events1;
 --1
 SELECT
 	user_id,
 	event_time,
 	event_type,
 	count() OVER(PARTITION BY user_id ORDER BY event_time) AS com_users
 FROM
 	mnoskov.events1;
 --2
 SELECT
 	user_id,
 	event_time,
 	event_type,
 	row_number() OVER (PARTITION BY user_id ORDER BY event_time) rn,
 	dense_rank() OVER (PARTITION BY user_id ORDER BY event_time) dr,
 	count() OVER(PARTITION BY user_id ORDER BY event_time) AS c,
 	sum(1) OVER(PARTITION BY user_id ORDER BY event_time) AS s
 FROM
 	mnoskov.events1;
 --3
 SELECT
 	user_id,
 	event_time,
 	event_type,
 	count() OVER(PARTITION BY user_id ORDER BY event_time ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW) AS com_users,
 	count() OVER(PARTITION BY user_id ORDER BY event_time ROWS BETWEEN UNBOUNDED PRECEDING AND UNBOUNDED FOLLOWING) AS com_users2
 FROM
 	mnoskov.events1;
 -- все четенько работает, никаких ошибкок нет
 --4-5-6
 WITH recent_events AS (
 	SELECT
 		user_id,
 		event_time,
 		event_type
 	FROM
 		mnoskov.events1
 	WHERE
 		event_time < now() - INTERVAL 1 DAY
 )
 SELECT
 	user_id,
 	count(*) AS cnt_user_actions
 FROM
 	recent_events
 GROUP BY
 	user_id
 ORDER BY
 	cnt_user_actions DESC
 LIMIT
 	5;
 -- не понял, как сделать 6 пункт.
 -- Ответ на вопрос - ну да - поддерживает.