[
https://issues.apache.org/jira/browse/FLINK-11427?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17108072#comment-17108072
]
Jingsong Lee commented on FLINK-11427:
--------------------------------------
I'm not sure if we need to add this code. Because it will bring a lot of
dependencies (parquet-proto) to the parquet project, which will not be used
most of the time.
Considering the simplicity of the code, I think it's enough to just add
documentation.
{code:java}
public static <T extends Message> ParquetWriterFactory create(Class<T> type) {
class Builder extends ParquetWriter.Builder<T, Builder> {
private Builder(OutputFile path) {
super(path);
}
@Override
protected Builder self() {
return this;
}
@Override
protected WriteSupport<T> getWriteSupport(Configuration conf) {
return new ProtoWriteSupport<>(type);
}
}
return new ParquetWriterFactory<>(out -> new Builder(out).build());
}{code}
Just add this example to documentation.
> Protobuf parquet writer implementation
> --------------------------------------
>
> Key: FLINK-11427
> URL: https://issues.apache.org/jira/browse/FLINK-11427
> Project: Flink
> Issue Type: Improvement
> Components: Formats (JSON, Avro, Parquet, ORC, SequenceFile)
> Reporter: Guang Hu
> Assignee: Yun Gao
> Priority: Trivial
> Labels: pull-request-available, usability
> Fix For: 1.11.0
>
> Time Spent: 10m
> Remaining Estimate: 0h
>
> Right now only ParquetAvroWriters exist to create ParquetWriterFactory. We
> want to implement a protobuf ParquetProtoWriters to create
> ParquetWriterFactory. I am happy to submit a PR if this approach sounds good
> .
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
