StarRocks source connector

Description

Read external data source data through StarRocks. The internal implementation of StarRocks source connector is obtains the query plan from the frontend (FE), delivers the query plan as a parameter to BE nodes, and then obtains data results from BE nodes.

Key features

• batch
• stream
• exactly-once
• schema projection
• parallelism
• support user-defined split

Options

node_urls [list]

StarRocks cluster address, the format is ["fe_ip:fe_http_port", ...]

username [string]

StarRocks user username

password [string]

StarRocks user password

database [string]

The name of StarRocks database

table [string]

The name of StarRocks table

scan_filter [string]

Filter expression of the query, which is transparently transmitted to StarRocks. StarRocks uses this expression to complete source-side data filtering.

e.g.

"tinyint_1 = 100"

schema [config]

fields [Config]

The schema of the starRocks that you want to generate

e.g.

schema {
    fields {
        name = string
        age = int
    }
  }

request_tablet_size [int]

The number of StarRocks Tablets corresponding to an Partition. The smaller this value is set, the more partitions will be generated. This will increase the parallelism on the engine side, but at the same time will cause greater pressure on StarRocks.

The following is an example to explain how to use request_tablet_size to controls the generation of partitions

the tablet distribution of StarRocks table in cluster as follower
be_node_1 tablet[1, 2, 3, 4, 5]
be_node_2 tablet[6, 7, 8, 9, 10]
be_node_3 tablet[11, 12, 13, 14, 15]
1.If not set request_tablet_size, there will no limit on the number of tablets in a single partition. The partitions will be generated as follows  
partition[0] read data of tablet[1, 2, 3, 4, 5] from be_node_1 
partition[1] read data of tablet[6, 7, 8, 9, 10] from be_node_2 
partition[2] read data of tablet[11, 12, 13, 14, 15] from be_node_3 
2.if set request_tablet_size=3, the limit on the number of tablets in a single partition is 3. The partitions will be generated as follows
partition[0] read data of tablet[1, 2, 3] from be_node_1 
partition[1] read data of tablet[4, 5] from be_node_1 
partition[2] read data of tablet[6, 7, 8] from be_node_2 
partition[3] read data of tablet[9, 10] from be_node_2 
partition[4] read data of tablet[11, 12, 13] from be_node_3 
partition[5] read data of tablet[14, 15] from be_node_3

scan_connect_timeout_ms [int]

requests connection timeout sent to StarRocks

scan_query_timeout_sec [int]

Query the timeout time of StarRocks, the default value is 1 hour, -1 means no timeout limit

scan_keep_alive_min [int]

The keep-alive duration of the query task, in minutes. The default value is 10. we recommend that you set this parameter to a value greater than or equal to 5.

scan_batch_rows [int]

The maximum number of data rows to read from BE at a time. Increasing this value reduces the number of connections established between engine and StarRocks and therefore mitigates overhead caused by network latency.

scan_mem_limit [long]

The maximum memory space allowed for a single query in the BE node, in bytes. The default value is 2147483648 (2 GB).

max_retries [int]

number of retry requests sent to StarRocks

scan.params. [string]

The parameter of the scan data from be

Example

source {
  StarRocks {
    nodeUrls = ["starrocks_e2e:8030"]
    username = root
    password = ""
    database = "test"
    table = "e2e_table_source"
    scan_batch_rows = 10
    max_retries = 3
    fields {
       BIGINT_COL = BIGINT
       LARGEINT_COL = STRING
       SMALLINT_COL = SMALLINT
       TINYINT_COL = TINYINT
       BOOLEAN_COL = BOOLEAN
       DECIMAL_COL = "DECIMAL(20, 1)"
       DOUBLE_COL = DOUBLE
       FLOAT_COL = FLOAT
       INT_COL = INT
       CHAR_COL = STRING
       VARCHAR_11_COL = STRING
       STRING_COL = STRING
       DATETIME_COL = TIMESTAMP
       DATE_COL = DATE
    }
    scan.params.scanner_thread_pool_thread_num = "3"
  }
}

Changelog

next version

• Add StarRocks Source Connector