Apache Hadoop Integration

Parquet Export Supports S3

Using EXPORT TO PARQUET, you can now export data to S3 in addition to HDFS and the local disk. For more information, see Exporting to S3.

HCatalog Connector Supports Custom Hive Partitions

By default Hive stores partition information under the path for the table definition, which might be local to Hive. However, a Hive user can choose to store partition information elsewhere, such as in a shared location like S3. In previous versions, Vertica did not include custom partitions in its queries. Now, during query execution Vertica can query Hive for this information. Because the additional Hive queries can be expensive, you specify when creating the HCatalog schema whether to look for custom partitions.

For more information, see Defining a Schema Using the HCatalog Connector.