Jeffrey Carter created FLINK-9126:
-------------------------------------
Summary: Ability for CassandraInputFormat to output data into a
Custom Cassandra Annotated Pojo
Key: FLINK-9126
URL: https://issues.apache.org/jira/browse/FLINK-9126
Project: Flink
Issue Type: New Feature
Components: DataSet API
Affects Versions: 1.4.2
Reporter: Jeffrey Carter
Fix For: 1.5.0
Attachments: CassandraPojoInputFormatText.rtf
*First time proposing new update so apologies if I missed anything*
Currently the DataSet API only has the ability to output data received from
Cassandra as a source in as a Tuple. This would be allow the data to be output
as a custom POJO that the user has created that has been annotated using
Datastax API. This would remove the need of very long Tuples to be created by
the DataSet and then mapped to the custom POJO.
The changes to the CassandraInputFormat object would be minimal, but would
require importing the Datastax API into the class. Another option is to make a
similar, but slightly different class called CassandraPojoInputFormat.
I have already gotten code for this working in my own project, but want other
thoughts as to the best way this should go about being implemented.
//Example of its use in main
CassandraPojoInputFormat<CustomCassandraPojo> cassandraInputFormat = new
CassandraPojoInputFormat<>(queryToRun, defaultClusterBuilder,
CustomCassandraPojo.class);
cassandraInputFormat.configure(null);
cassandraInputFormat.open(null);
DataSet<CustomCassandraPojo> outputTestSet =
exEnv.createInput(cassandraInputFormat, TypeInformation.of(new
TypeHint<CustomCassandraPojo>(){}));
//The class that I currently have set up
[^CassandraPojoInputFormatText.rtf]
Will make another Jira Issue for the Output version next if this is approved
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
