[feature] HDFS interface

[steremma]

I have started to work on providing a new datasource, targeting semi structured data residing on a distributed file system (HDFS/S3) such as:

• csv
• json
• xml

I have started to work on that by designing a spark API imitating the functionality offered by the already existing datasources.

I would be interested to know if people would appreciate the addition of such a feature.
I also have some ideas regarding the implementation, specifically I am now adding functionality in a new
module under connectors/

Any input is valuable at this point!

[mistercrunch]

You'll need an engine that can deserialize, aggregate & filter data as you implement the interface, and Superset isn't the place to do that. Depending on your serde you may want to write a deserializer for Presto or something like Apache Drill.

[mistercrunch]

Notice: this issue has been closed because it has been inactive for 392 days. Feel free to comment and request for this issue to be reopened.