|| at Jul 1, 2013 at 3:54 pm
Impala has the advantage of being able to query data stored in a wide
variety of file formats. Some of these are row based (like sequence files
or delimited text files) and some are columnar (like Parquet).
You can convert between formats by reading them from one table and writing
to another which has a different format. Impala currently supports writing
to Parquet tables and text-delimited tables.
On 1 July 2013 03:42, wrote:
Impala uses hive metastore. If a hive table is loaded and queried by
impala,is there any kind of data migration or conversion from hive format
to columnar format.I mean what my understanding is impala uses data in
columnar format.If yes then at what point is this data is converted to