ClickHouse/docs/en/engines/table-engines/integrations/redis.md

slug	sidebar_position	sidebar_label
/en/engines/table-engines/integrations/redis	43	Redis

Redis

This engine allows integrating ClickHouse with Redis. For Redis takes kv model, we strongly recommend you only query it in a point way, such as where k=xx or where k in (xx, xx).

Creating a Table

CREATE TABLE [IF NOT EXISTS] [db.]table_name
(
    name1 [type1],
    name2 [type2],
    ...
) ENGINE = Redis(host:port[, db_index[, password[, pool_size]]]) PRIMARY KEY(primary_key_name);

Engine Parameters

• host:port — Redis server address, you can ignore port and default Redis port 6379 will be used.

• db_index — Redis db index range from 0 to 15, default is 0.

• password — User password, default is blank string.

• pool_size — Redis max connection pool size, default is 16.

• primary_key_name - any column name in the column list.

• primary must be specified, it supports only one column in the primary key. The primary key will be serialized in binary as a Redis key.

• columns other than the primary key will be serialized in binary as Redis value in corresponding order.

• queries with key equals or in filtering will be optimized to multi keys lookup from Redis. If queries without filtering key full table scan will happen which is a heavy operation.

Usage Example

Create a table in ClickHouse which allows to read data from Redis:

CREATE TABLE redis_table
(
    `k` String,
    `m` String,
    `n` UInt32
)
ENGINE = Redis('redis1:6379') PRIMARY KEY(k);

Insert:

INSERT INTO redis_table Values('1', 1, '1', 1.0), ('2', 2, '2', 2.0);

Query:

SELECT COUNT(*) FROM redis_table;

┌─count()─┐
│       2 │
└─────────┘

SELECT * FROM redis_table WHERE key='1';

┌─key─┬─v1─┬─v2─┬─v3─┐
│ 1   │  1 │ 1  │  1 │
└─────┴────┴────┴────┘

SELECT * FROM redis_table WHERE v1=2;

┌─key─┬─v1─┬─v2─┬─v3─┐
│ 2   │  2 │ 2  │  2 │
└─────┴────┴────┴────┘

Update:

Note that the primary key cannot be updated.

ALTER TABLE redis_table UPDATE v1=2 WHERE key='1';

Delete:

ALTER TABLE redis_table DELETE WHERE key='1';

Truncate:

Flush Redis db asynchronously. Also Truncate support SYNC mode.

TRUNCATE TABLE redis_table SYNC;

Limitations

Redis engine also supports scanning queries, such as where k > xx, but it has some limitations:

1. Scanning query may produce some duplicated keys in a very rare case when it is rehashing. See details in Redis Scan
2. During the scanning, keys could be created and deleted, so the resulting dataset can not represent a valid point in time.