PXF: PXF is an extensible framework that allows HAWQ to query external system data
Let's learn Query federation
This topic describes how to access Hive data using PXF. Link
Previously, in order to query Hive tables using HAWQ and PXF, you needed to create an external table in PXF that described the target table's Hive metadata. Since HAWQ is now integrated with HCatalog, HAWQ can use metadata stored in HCatalog instead of external tables created for PXF. HCatalog is built on top of the Hive metastore and incorporates Hive's DDL. This provides several advantages:
You do not need to know the table schema of your Hive tables
You do not need to manually enter information about Hive table location or format
If Hive table metadata changes, HCatalog provides updated metadata. This is in contrast to the use of static external PXF tables to define Hive table metadata for HAWQ.
HAWQ retrieves table metadata from HCatalog using PXF.
HAWQ creates in-memory catalog tables from the retrieved metadata. If a table is referenced multiple times in a transaction, HAWQ uses its in-memory metadata to reduce external calls to HCatalog.
PXF queries Hive using table metadata that is stored in the HAWQ in-memory catalog tables. Table metadata is dropped at the end of the transaction.