如需在 ByteHouse 中删除数据，可通过 ALTER DELETE 方式，更多操作详情请参考删除操作（DELETE）。

语法 #

ALTER TABLE [tableIdentifier] DELETE [IN PARTITION partition_id] WHERE filter_expr;

参数说明：
IN PARTITION partition_id：可选，如果表数据量较大，或者表的 part 数量较多，建议添加该条件，分区级别执行 delete 操作，更快、更稳定。

示例 #

-- 1. 创建库
CREATE DATABASE IF NOT EXISTS sample_db ON CLUSTER sample_cluster;

-- 2. 创建测试表
DROP TABLE IF EXISTS sample_db.sample_table_local ON CLUSTER sample_cluster;
DROP TABLE IF EXISTS sample_db.sample_table ON CLUSTER sample_cluster;
CREATE TABLE IF NOT EXISTS sample_db.sample_table_local ON CLUSTER sample_cluster
(
    `order_by_column` String,
    `partition_column` Date,
    `browser` String,
    `hour` Int32,
    `x` Int32,
    `p` Int32
) ENGINE = HaMergeTree('/clickhouse/tables/{shard}/{database}/local_table_name', '{replica}')
PARTITION BY partition_column
ORDER BY (`order_by_column`);

CREATE TABLE sample_db.sample_table
    AS sample_db.sample_table_local
    ENGINE = Distributed(sample_cluster, sample_db, sample_table_local, hour);

-- 3. 插入数据
INSERT INTO sample_db.sample_table_local ON CLUSTER sample_cluster
VALUES
('order1', '2020-11-21', 'chrome', 8, 10, 2),
('order2', '2020-11-02', 'chrome', 8, 10, 2),
('order3', '2020-10-05', 'firefox', 9, 20, 2);

-- 4. 删除数据
ALTER TABLE sample_db.sample_table_local ON CLUSTER sample_cluster 
    DELETE WHERE browser = 'firefox';

-- 5.查询删除是否成功
SELECT * 
FROM clusterAllReplicas(sample_cluster, system.mutations) 
WHERE 
database = 'sample_db' 
AND table = 'sample_table_local';