Github user fhueske commented on a diff in the pull request:
https://github.com/apache/flink/pull/1941#discussion_r63174853
--- Diff:
flink-batch-connectors/flink-jdbc/src/main/java/org/apache/flink/api/java/io/jdbc/JDBCInputFormat.java
---
@@ -157,14 +216,25 @@ public boolean reachedEnd() throws IOException {
* @throws java.io.IOException
*/
@Override
- public OUT nextRecord(OUT tuple) throws IOException {
+ public Row nextRecord(Row row) throws IOException {
try {
- resultSet.next();
- if (columnTypes == null) {
- extractTypes(tuple);
+ hasNext = resultSet.next();
+ if (!hasNext) {
+ return null;
+ }
+ try {
+ //This throws a NPE when the TypeInfo is not
passed to the InputFormat,
+ //i.e. KryoSerializer used to generate the
passed row
+ row.productArity();
--- End diff --
OK, I see. How about we extend the InputFormat to implement the
`ResultTypeQueryable` interface and let users either specify all field types or
at least the number of result attributes via the `JDBCInputFormatBuilder`.
Then we do not have to fall back to the KryoSerializer that does creates
corrupt `Row` objects and users do not have to specify types in the
`env.createInput()` method.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
