蜂巢中具有分区的外部表

Question

I have a bunch of tsv files in HDFS in a directory structure that follows the partition convention where an event_dt is the partition.

some_path/event_dt=2017-04-30
some_path/event_dt=2017-05-01

and so on.

The issue is that event_dt is also one of the columns. The second one in particular. But I cannot specify so since event_dt cannot appear in the table schema and in the PARTITIONED BY statement. That triggers:

 Column repeated in partitioning columns

Is there a way around this other than using different names. It is, after all, the same information.

Answer 1

3 options if you dont want to rename the column.

1. If your event_dt is the last column in your csv, you create the table excluding this column.
2. During the ingestion process exclude this information of your data, transforming the data from one place to another where the target table is partitioned by even_dt (not the most efficient way)
3. create a view on top of your table excluding one of the columns, anyway the original table will need the rename .