Common table expression CTE in Clickhouse

What is a common table expression (CTE) ？ In this paper , Learn how to ClickHouse Use in the database CTE, And track use cases through examples

Use when CTE Very convenient :

• When a request can get data , And its size is suitable for memory space
• The result of this query needs to be used more than once
• Create recursive queries

The additional benefit is to improve SQL Readability of queries .

CTE What is the difference between temporary tables and nested queries ？

• If the subquery is relevant , Then its call will repeat for each line in the selection , This significantly increases the cost of executing this query .
• Filling a temporary table with a large amount of data will cause load on the disk .
• Due to the particularity of storing temporary tables , Using temporary tables to execute queries will increase execution time .

grammar

ClickHouse Support both WITH< expression > AS < identifier > , Also support WITH< identifier > AS < Subquery expression > grammar .

• Use WITH initialization CTE.
• Provide a name for the query .
• Next is AS.
• Define query
• If more than one is needed CTE, Please separate them with commas .

WITH locations AS
    (
        SELECT location
        FROM table
        WHERE date > (today() - 10)
    )
SELECT *
FROM locations

• Use WITH initialization CTE.
• Define an expression
• Next is AS.
• Give the expression a name .
• If more than one is needed CTE, Please separate them with commas .

WITH ('USA', 'BRA') AS locations
SELECT 'ARG' IN (locations)

Example

Build table :

CREATE TABLE SpareParts
(
    `id` UInt32,
    `partName` String,
    `partOrigin` String,
    `storeID` UInt32
)
ENGINE = MergeTree()
ORDER BY id

Insert :

INSERT INTO SpareParts VALUES (1, 'headlight', 'USA', 1)
INSERT INTO SpareParts VALUES (2, 'hood', 'JPN', 1)
INSERT INTO SpareParts VALUES (3, 'bumper', 'USA', 1)
INSERT INTO SpareParts VALUES (4, 'radiator', 'BRA', 3)
INSERT INTO SpareParts VALUES (5, 'wheel', 'BRA', 2)
INSERT INTO SpareParts VALUES (6, 'stabilizer', 'ARG', 3)
INSERT INTO SpareParts VALUES (7, 'absorber', 'TUR', 2)
INSERT INTO SpareParts VALUES (8, 'cable', 'MEX', 1)
INSERT INTO SpareParts VALUES (9, 'spring', 'MEX', 3)
INSERT INTO SpareParts VALUES (10, 'door', 'USA', 2)

choice :

WITH
    originsByStore AS
    (
        SELECT
            storeID,
            groupArray(partOrigin) AS origins
        FROM SpareParts
        GROUP BY storeID
    ),
    partsByStore AS
    (
        SELECT
            storeID,
            groupArray(partName) AS partNames
        FROM SpareParts
        GROUP BY storeID
    ),
    has(origins, 'USA') = 1 AS isUSA
SELECT
    storeID,
    origins,
    partNames,
    isUSA
FROM originsByStore AS t1
LEFT JOIN
(
    SELECT
        storeID,
        partNames
    FROM partsByStore
) AS t2 USING (storeID)

result :

Blogger's personal blog website ：https://www.beierblog.com

The original author of this article ： Whimsical school 、 A programmer who works hard to share .

The platform for the first article ： WeChat official account 【 Programming talent 】

Originality is not easy. ！ If you think the article is good , Let's pay attention to the official account , Conduct give the thumbs-up （ Looking at ）、 forward Three companies walk up ！ Thank you. ！