Skip to main content
Version: Next

ClickhouseFile

Clickhouse file sink connector

Description

Generate the clickhouse data file with the clickhouse-local program, and then send it to the clickhouse server, also call bulk load. This connector only support clickhouse table which engine is 'Distributed'.And internal_replication option should be true. Supports Batch and Streaming mode.

Key features

tip

Write data to Clickhouse can also be done using JDBC

Options

NameTypeRequiredDefault
hoststringyes-
databasestringyes-
tablestringyes-
usernamestringyes-
passwordstringyes-
clickhouse_local_pathstringyes-
sharding_keystringno-
copy_methodstringnoscp
node_free_passwordbooleannofalse
node_passlistno-
node_pass.node_addressstringno-
node_pass.usernamestringno"root"
node_pass.passwordstringno-
compatible_modebooleannofalse
file_fields_delimiterstringno"\t"
file_temp_pathstringno"/tmp/seatunnel/clickhouse-local/file"
common-optionsno-

host [string]

ClickHouse cluster address, the format is host:port , allowing multiple hosts to be specified. Such as "host1:8123,host2:8123" .

database [string]

The ClickHouse database

table [string]

The table name

username [string]

ClickHouse user username

password [string]

ClickHouse user password

sharding_key [string]

When ClickhouseFile split data, which node to send data to is a problem, the default is random selection, but the 'sharding_key' parameter can be used to specify the field for the sharding algorithm.

clickhouse_local_path [string]

The address of the clickhouse-local program on the spark node. Since each task needs to be called, clickhouse-local should be located in the same path of each spark node.

copy_method [string]

Specifies the method used to transfer files, the default is scp, optional scp and rsync

node_free_password [boolean]

Because seatunnel need to use scp or rsync for file transfer, seatunnel need clickhouse server-side access. If each spark node and clickhouse server are configured with password-free login, you can configure this option to true, otherwise you need to configure the corresponding node password in the node_pass configuration

node_pass [list]

Used to save the addresses and corresponding passwords of all clickhouse servers

node_pass.node_address [string]

The address corresponding to the clickhouse server

node_pass.username [string]

The username corresponding to the clickhouse server, default root user.

node_pass.password [string]

The password corresponding to the clickhouse server.

compatible_mode [boolean]

In the lower version of Clickhouse, the ClickhouseLocal program does not support the --path parameter, you need to use this mode to take other ways to realize the --path parameter function

file_fields_delimiter [string]

ClickhouseFile uses csv format to temporarily save data. If the data in the row contains the delimiter value of csv, it may cause program exceptions. Avoid this with this configuration. Value string has to be an exactly one character long

file_temp_path [string]

The directory where ClickhouseFile stores temporary files locally.

common options

Sink plugin common parameters, please refer to Sink Common Options for details

Examples

ClickhouseFile {
host = "192.168.0.1:8123"
database = "default"
table = "fake_all"
username = "default"
password = ""
clickhouse_local_path = "/Users/seatunnel/Tool/clickhouse local"
sharding_key = "age"
node_free_password = false
node_pass = [{
node_address = "192.168.0.1"
password = "seatunnel"
}]
}

Changelog

2.2.0-beta 2022-09-26

  • Support write data to ClickHouse File and move to ClickHouse data dir

Next version

  • [BugFix] Fix generated data part name conflict and improve file commit logic 3416
  • [Feature] Support compatible_mode compatible with lower version Clickhouse 3416