Danny Chan

The contract here is: the partition **field** shoud be in the table schema anyway.

cc @the-other-tim-brown for the code reviewing.

There is no guarantee that `#getInsertValue` should be invoked after `#combineAndGetUpdateValue`, actually we need the value returned by `#getInsertValue` to hand it over to `#combineAndGetUpdateValue`.

> And why does it only happen with Merge On Read My guess is we use the `#getInsertValue` to deserialize the payload from logs. Maybe I lost some reminisenses and...

@KnightChess do you have intreast to push-forward this feature?

Did you try the Hudi 0.14.1 release? Did you enable the metadata table?

```java Caused by: java.util.NoSuchElementException: FileID xxxxx of partition path dt=2019-02-20 does not exist. at org.apache.hudi.io.HoodieMergeHandle.getLatestBaseFile(HoodieMergeHandle.java:159) at org.apache.hudi.io.HoodieMergeHandle.(HoodieMergeHandle.java:121) at org.apache.hudi.io.FlinkMergeHandle.(FlinkMergeHandle.java:70) at org.apache.hudi.io.FlinkConcatHandle.(FlinkConcatHandle.java:53) at org.apache.hudi.client.HoodieFlinkWriteClient.getOrCreateWriteHandle(HoodieFlinkWriteClient.java:557) at org.apache.hudi.client.HoodieFlinkWriteClient.insert(HoodieFlinkWriteClient.java:175) at org.apache.hudi.sink.StreamWriteFunction.lambda$initWriteFunction$0(StreamWriteFunction.java:181) at org.apache.hudi.sink.StreamWriteFunction.lambda$flushRemaining$7(StreamWriteFunction.java:461) ```...

> I actually see hudi could set many spark relate configures in SparkConf, most of them are related to parquet reader/writer. Are these options configurable?

Not really, you can create a JIRA to trace it.

Thanks for the contribution @waywtdcc , can you explain in high level how the hudi table is loaded and what is the refresh strategy of the table ?