Description
PXF is excellent at allowing hawq to do data federation queries.
With a JDBC PXF (https://github.com/kojec/pxf-field/tree/master/jdbc-pxf-ext - thanks kopec) we can connect to almost every external DB that we can imagine, and either at query time join data from within Hawq with our external Oracle DW or Mongo DB's without an expensive ETL process to bring that data to Hawq/Hadoop.
Likewise if we want faster query, we can clone the data in from that source very easily all with SQL still within Hawq, still a lot simpler than a full blown ETL process.
The downside of this atm is that we have to create an external table for each and every table.
The new native hcatalog feature very nicely shows that this is possible and this for hive tables, but is baked into hawq.
It would be great PXF could be extended where by we can register/create an EXTERNAL DATABASE where then hawq delegates to PXF very much like for the native hcatalog seems to.
To support this PXF would need extend support api's to return schema, tables and then field definitions.
Attachments
Issue Links
- relates to
-
HAWQ-459 Enhance Metadata Rest API to support table(s)
-
- Resolved
-