Greenplum Sink
1、Introduce
Greenplum sink
2、Version Support
Greenplum 5 and above
3、Connector name
Type | Name |
---|---|
Sync | greenplumsink、greenplumwriter |
--- | --- |
SQL | greenplum-x |
4、Parameter description
1、Sync
connection
- Description:param for Database connection,including jdbcUrl、schema、table and so on
- Required:required
- Type:List
- Default:none
"connection": [{
"jdbcUrl": ["jdbc:pivotal:greenplum://localhost:5432;DatabaseName=postgres"],
"table": ["table"],
"schema":"public"
}]
jdbcUrl
- Description:jdbc connection string for relational database,jdbcUrl reference:greenplum
- Required: required
- Type: string
- Default: none
schema
- Description:Database schema
- Required:optional
- Type:string
- Default:none
table
- Description:The table name of the target table. Currently only single table is supported, and multiple tables will be supported in the future
- Required:required
- Type:List
- Default:none
username
- Description:user name
- Required:required
- Type:String
- Default:none
password
- Description:password
- Required:required
- Type:String
- Default:none
column
- Description:The fields that the destination table needs to write data into, and the fields are separated by commas.for example: "column": ["id","name","age"]
- Required:required
- Type:List
- Default:none
fullcolumn
- Description:All fields in the destination table ,which is separated by English commas.for example: "column": ["id","name","age","hobby"],if not configured, it will be obtained in the system table
- Required:optional
- Type:List
- Default:none
preSql
- Description:the sql executed before writing data into the destination table
- Required:optional
- Type:List
- Default:none
postSql
- Description:the sql executed after writing data into the destination table
- Required:optional
- Type:List
- Default:none
writeMode
- Description:Controls writing data to the target table using insert into or replace into or ON DUPLICATE KEY UPDATE
- Required:required
- All options:insert/replace/update
- Type:String
- Default:insert
batchSize
- Description:The number of records submitted in batch at one time. This value can greatly reduce the number of network interactions between flinkx and the database and improve the overall throughput,Setting this value too large may cause the flinkx process to run oom
- Required:optional
- Type:int
- Default:1024
updateKey
- Description:When the write mode is update, you need to specify the value of this parameter as the unique index field
- attention:
- If this parameter is empty and the write mode is update, the application will automatically obtain the unique index in the database;
- If the data table does not have a unique index, but the required write mode is configured as update and, the application will write data in the way of insert;
- Required:optional
- Type:Map<String,List>
- for example:"updateKey": {"key": ["id"]}
- Default:none
semantic
- Description:sink operator support phase two commit
- attention: -If this parameter is blank, phase two commit is not enabled by default,which means sink operators do not support exactly-once semantics -Currently only supported exactly-once and at-least-once
- Required:optional
- Type:String
- for example:"semantic": "exactly-once"
- Default:at-least-once
2、SQL
connector
- Description:greenplum-x
- Required:required
- Type:String
- Default:none
url
- Description:jdbc:pivotal:greenplum://localhost:5432;DatabaseName=postgres
- Required:required
- Type:String
- Default:none
schema
- Description:Database schema
- Required:optional
- Type:string
- Default:none
table-name
- Description: table name
- Required:required
- Type:String
- Default:none
username
- Description:username
- Required:required
- Type:String
- Default:none
password
- Description:password
- Required:required
- Type:String
- Default:none
sink.buffer-flush.max-rows
- Description:Number of data pieces written in batch,Unit:piece.
- Required:optional
- Type:String
- Default:1024
sink.buffer-flush.interval
- Description:Batch write interval,Unit: ms
- Required:optional
- Type:String
- Default:10000
sink.all-replace
- Description:whether to replace all data in the database(If the original value in the database is not null, the new value is null, if it is true, it will be replaced with null)
- Required:optional
- Type:String
- Default:false
sink.parallelism
- Description:the parallelism of sink operator
- Required:optional
- Type:String
- Default:none
sink.semantic
- Description:sink operator support phase two commit
- attention: -If this parameter is blank, phase two commit is not enabled by default,which means sink operators do not support exactly-once semantics; -Currently only supported exactly-once and at-least-once
- Required:optional
- Type:String
- for example:"semantic": "exactly-once"
- Default:at-least-once
五、数据类型
Whether to support | Data Type |
---|---|
Supported | SMALLINT、SMALLSERIAL、INT2、INT、INTEGER、SERIAL、INT4、BIGINT、BIGSERIAL、OID、INT8、REAL、FLOAT4、FLOAT、DOUBLE PRECISION、FLOAT8、DECIMAL、NUMERIC、 CHARACTER VARYING、VARCHAR、CHARACTER、CHAR、TEXT、NAME、BPCHAR、BYTEA、TIMESTAMP、TIMESTAMPTZ、DATE、TIME、TIMETZ、 BOOLEAN、BOOL |
Unsupported | ARRAY etc. |
6. Example
The details are in flinkx-examples dir.