Skip to content
ClickHouse Docs
ClickHouse DocsClickHouse Docs
49.6kGet Started

REPLACE TABLE

View as MarkdownEdit this pageReport an issueSet up docs MCP

Overview

The REPLACE statement allows you to update a table atomically.

Ordinarily, if you need to delete some data from a table, you can create a new table and fill it with a SELECT statement that does not retrieve unwanted data, then drop the old table and rename the new one. This approach is demonstrated in the example below:

CREATE TABLE myNewTable AS myOldTable;

INSERT INTO myNewTable
SELECT * FROM myOldTable
WHERE CounterID <12345;

DROP TABLE myOldTable;

RENAME TABLE myNewTable TO myOldTable;

Instead of the approach above, it is also possible to use REPLACE (given you are using the default database engines) to achieve the same result:

REPLACE TABLE myOldTable
ENGINE = MergeTree()
ORDER BY CounterID
AS
SELECT * FROM myOldTable
WHERE CounterID <12345;

Syntax

{CREATE [OR REPLACE] | REPLACE} TABLE [db.]table_name

Examples

Consider the following table:

CREATE DATABASE base
ENGINE = Atomic;

CREATE OR REPLACE TABLE base.t1
(
    n UInt64,
    s String
)
ENGINE = MergeTree
ORDER BY n;

INSERT INTO base.t1 VALUES (1, 'test');

SELECT * FROM base.t1;

┌

We can use the REPLACE statement to clear all the data:

CREATE OR REPLACE TABLE base.t1
(
    n UInt64,
    s Nullable(String)
)
ENGINE = MergeTree
ORDER BY n;

INSERT INTO base.t1 VALUES (2, null);

SELECT * FROM base.t1;

┌

Or we can use the REPLACE statement to change the table structure:

REPLACE TABLE base.t1 (n UInt64)
ENGINE = MergeTree
ORDER BY n;

INSERT INTO base.t1 VALUES (3);

SELECT * FROM base.t1;

┌
Navigation