[
https://issues.apache.org/jira/browse/FLINK-16622?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17060616#comment-17060616
]
gkgkgk commented on FLINK-16622:
--------------------------------
[~jark]Yes, I noticed this usage. But I'm used to define the schema explicitly.
And, I found that this error is affected by the following codeļ¼
@SuppressWarnings("unchecked")
public static <E> TypeInformation<E[]> OBJECT_ARRAY(TypeInformation<E>
elementType) {
if (elementType == Types.STRING) {
return (TypeInformation)
BasicArrayTypeInfo.STRING_ARRAY_TYPE_INFO;
}
return ObjectArrayTypeInfo.getInfoFor(elementType);
}
When I commented out the first three lines of code, it worked fine, and I
tested json-format, csv-format
> Type ARRAY<STRING> of table field 'id' does not match with the physical type
> LEGACY('ARRAY', 'ANY<[Ljava.lang.String;, rO0ABXNyAD......>') of the 'id'
> field of the TableSource return type.
> --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
>
> Key: FLINK-16622
> URL: https://issues.apache.org/jira/browse/FLINK-16622
> Project: Flink
> Issue Type: Bug
> Components: Table SQL / API
> Affects Versions: 1.10.0
> Reporter: gkgkgk
> Priority: Major
>
> when i define a ARRAY<VARCHAR> field in table schema and define field in
> 'format.json-schema' like this
> 'format.json-schema' = '{
> "type": "object",
> "properties": {
> "id": {
> "type": "array",
> "items":{"type":"string"}
> }
> }'
> there is a error flow this:
> Exception in thread "main" org.apache.flink.table.api.ValidationException:
> Type ARRAY<STRING> of table field 'id' does not match with the physical type
> LEGACY('ARRAY', 'ANY<[Ljava.lang.String;, rO0ABXNyADdvc......>') of the 'id'
> field of the TableSource return type.
> the flow is my ddl:
> --source
> CREATE TABLE dwd_user_log (
> id ARRAY<VARCHAR>,
> ctime TIMESTAMP(3),
> pageId VARCHAR,
> deviceId VARCHAR,
> WATERMARK FOR ctime AS ctime - INTERVAL '10' SECOND
> ) WITH (
> 'connector.type' = 'kafka',
> 'connector.version' = 'universal',
> 'connector.topic' = 'dev_dwd_user_log_02',
> 'connector.startup-mode' = 'latest-offset',
> 'connector.properties.0.key' = 'zookeeper.connect',
> 'connector.properties.0.value' = 'node14.example.com:2181',
> 'connector.properties.1.key' = 'bootstrap.servers',
> 'connector.properties.1.value' = 'node14.example.com:9092',
> 'connector.properties.2.key' = 'group.id',
> 'connector.properties.2.value' = 'dev-group',
> 'update-mode' = 'append',
> 'format.type' = 'json',
> -- 'format.derive-schema' = 'true'
> 'format.json-schema' = '{
> "type": "object",
> "properties": {
> "id": {
> "type": "array",
> "items":{"type":"string"}
> },
> "ctime": {
> "type": "string",
> "format": "date-time"
> },
> "pageId": {
> "type": "string"
> },
> "deviceId": {
> "type": "string"
> }
> }
> }'
> -- 'schema.1.rowtime.timestamps.type' = 'from-field',
> -- 'schema.1.rowtime.timestamps.from' = 'ctime',
> -- 'schema.1.rowtime.watermarks.type' = 'periodic-bounded',
> -- 'schema.1.rowtime.watermarks.delay' = '10000'
> -- 'schema.1.from' = 'ctime'
> );
> -- sink
> -- sink for pv
> CREATE TABLE dws_pv (
> windowStart TIMESTAMP(3),
> windowEnd TIMESTAMP(3),
> pageId VARCHAR,
> id ARRAY<VARCHAR>,
> viewCount BIGINT
> ) WITH (
> 'connector.type' = 'kafka',
> 'connector.version' = 'universal',
> 'connector.topic' = 'dev_dws_pvuv_02',
> 'connector.startup-mode' = 'latest-offset',
> 'connector.properties.0.key' = 'zookeeper.connect',
> 'connector.properties.0.value' = 'node14.example.com:2181',
> 'connector.properties.1.key' = 'bootstrap.servers',
> 'connector.properties.1.value' = 'node14.example.com:9092',
> 'connector.properties.2.key' = 'group.id',
> 'connector.properties.2.value' = 'dev-group',
> 'update-mode' = 'append',
> 'format.type' = 'json',
> 'format.derive-schema' = 'true'
> );
> -- pv
> INSERT INTO dws_pv
> SELECT
> TUMBLE_START(ctime, INTERVAL '20' SECOND) AS windowStart,
> TUMBLE_END(ctime, INTERVAL '20' SECOND) AS windowEnd,
> pageId,
> id,
> COUNT(deviceId) AS viewCount
> FROM dwd_user_log
> GROUP BY TUMBLE(ctime, INTERVAL '20' SECOND),pageId,id;
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
