aljoscha commented on issue #9434: [FLINK-13634][Formats]: added module to 
compress data using StreamingFileSink
URL: https://github.com/apache/flink/pull/9434#issuecomment-559187210
 
 
   Thanks for the quick response!
   
   I'm afraid I didn't express myself too clearly. We can still have the 
builder pattern, i.e. `CompressionWriterFactory` can have the builder methods 
for setting extractor and codec. The naming of the `BulkWriter.Factory` is 
unfortunately not very good and it could have been called `BulkWriter.Builder` 
instead. I thought that an `Extractor` was always needed so the static method 
on `HadoopCompressionWriters` could have been `forExtractor(Extractor)` and 
that returns the factory where you have the methods for setting the compression 
codec. If that's not the case you can also have a static method without 
parameters and do everything via the "builder".
   
   Also, we shouldn't have both `hadoopcompress` and `hadoop`. We can either 
have only `hadoopcompress` or `compress` and then inside that 
`compress.hadoop`. Are you actually planning to add more compression codecs 
besides Hadoop? I think planning for those future additions would mean we have 
to potentially make the interfaces more independent of Hadoop.
   
   Some Javadoc comments are outdated now, for example 
`CompressionWriterFactory` has
   ```
    * A {@link BulkWriter} implementation that compress file.
   ```
   
   Sorry for the inconvenience for you!

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
[email protected]


With regards,
Apache Git Services

Reply via email to