[
https://issues.apache.org/jira/browse/FLINK-15326?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
hehuiyuan updated FLINK-15326:
------------------------------
Description:
Env :
Flink 1.9.1
table-planner-blink
Question:
If i have a kafka sink table with json schema:
{code:java}
String jsonSchema =
"{
type:'object',
properties:{
name: { type: 'string' },
age: { type: 'integer' },
sex: { type: 'string' }
}
}";
JsonRowDeserializationSchema deserializationSchema = new
JsonRowDeserializationSchema(jsonSchema.toString());
TypeInformation fieldTypes = deserializationSchema.getProducedType();
Kafka kafka = new Kafka.......
Schema schema = new Schema......
tableEnvironment.connect(kafka)
.withFormat( new Json().jsonSchema(jsonSchema))
.withSchema( schema )
.inAppendMode()
.registerTableSink("sink_example2")
;
String sinksql = "insert into sink_example2 select * from table2"
tableEnvironment.sqlUpdate(sinksql);
{code}
Error:
{code:java}
Query result schema: [name: String, age: BigDecimal, sex: String]
TableSink schema: [name: String, age: BigDecimal, sex: String]
{code}
The table `table2` : table schema
{code:java}
[2019-12-19 18:10:16,937] INFO t2: root
|-- name: STRING
|-- age: DECIMAL(10, 0)
|-- sex: STRING
{code}
When i use kafka to read data for json schema , i understand the integer type
in json is mapped DECIMAL(38, 18) in flink table
{code:java}
|-- name: STRING
|-- age: DECIMAL(38, 18)
|-- sex: STRING
{code}
That's why i know to set decimal precision
{code:java}
String sinksql = "insert into sink_example2 select name CAST(age as
decimal(38,18) ) as age, sex from table2"
{code}
JSON format needs to pay attention to this problem.
was:
Env :
Flink 1.9.1
table-planner-blink
Question:
If i have a kafka sink table with json schema:
{code:java}
String jsonSchema =
"{
type:'object',
properties:{
name: { type: 'string' },
age: { type: 'integer' },
sex: { type: 'string' }
}
}";
JsonRowDeserializationSchema deserializationSchema = new
JsonRowDeserializationSchema(jsonSchema.toString());
TypeInformation fieldTypes = deserializationSchema.getProducedType();
Kafka kafka = new Kafka.......
Schema schema = new Schema......
tableEnvironment.connect(kafka)
.withFormat( new Json().jsonSchema(jsonSchema))
.withSchema( schema )
.inAppendMode()
.registerTableSink("sink_table")
;
String sinksql = "insert into sink_example2 select * from table2"
tableEnvironment.sqlUpdate(sinksql);
{code}
Error:
{code:java}
Query result schema: [name: String, age: BigDecimal, sex: String]
TableSink schema: [name: String, age: BigDecimal, sex: String]
{code}
The table `table2` : table schema
{code:java}
[2019-12-19 18:10:16,937] INFO t2: root
|-- name: STRING
|-- age: DECIMAL(10, 0)
|-- sex: STRING
{code}
When i use kafka to read data for json schema , i understand the integer type
in json is mapped DECIMAL(38, 18) in flink table
{code:java}
|-- name: STRING
|-- age: DECIMAL(38, 18)
|-- sex: STRING
{code}
That's why i know to set decimal precision
{code:java}
String sinksql = "insert into sink_example2 select name CAST(age as
decimal(38,18) ) as age, sex from table2"
{code}
JSON format needs to pay attention to this problem.
> Add document description for DECIMAL(38, 18) when the sink table uses json
> schema
> ----------------------------------------------------------------------------------
>
> Key: FLINK-15326
> URL: https://issues.apache.org/jira/browse/FLINK-15326
> Project: Flink
> Issue Type: Wish
> Reporter: hehuiyuan
> Priority: Major
>
> Env :
> Flink 1.9.1
> table-planner-blink
>
> Question:
> If i have a kafka sink table with json schema:
>
> {code:java}
> String jsonSchema =
> "{
> type:'object',
> properties:{
> name: { type: 'string' },
> age: { type: 'integer' },
> sex: { type: 'string' }
> }
> }";
> JsonRowDeserializationSchema deserializationSchema = new
> JsonRowDeserializationSchema(jsonSchema.toString());
> TypeInformation fieldTypes = deserializationSchema.getProducedType();
> Kafka kafka = new Kafka.......
> Schema schema = new Schema......
> tableEnvironment.connect(kafka)
> .withFormat( new Json().jsonSchema(jsonSchema))
> .withSchema( schema )
> .inAppendMode()
> .registerTableSink("sink_example2")
> ;
> String sinksql = "insert into sink_example2 select * from table2"
> tableEnvironment.sqlUpdate(sinksql);
> {code}
>
>
> Error:
> {code:java}
> Query result schema: [name: String, age: BigDecimal, sex: String]
> TableSink schema: [name: String, age: BigDecimal, sex: String]
> {code}
> The table `table2` : table schema
>
> {code:java}
> [2019-12-19 18:10:16,937] INFO t2: root
> |-- name: STRING
> |-- age: DECIMAL(10, 0)
> |-- sex: STRING
>
> {code}
>
> When i use kafka to read data for json schema , i understand the integer type
> in json is mapped DECIMAL(38, 18) in flink table
>
> {code:java}
> |-- name: STRING
> |-- age: DECIMAL(38, 18)
> |-- sex: STRING
> {code}
>
> That's why i know to set decimal precision
> {code:java}
> String sinksql = "insert into sink_example2 select name CAST(age as
> decimal(38,18) ) as age, sex from table2"
> {code}
> JSON format needs to pay attention to this problem.
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
