3

I have to create a table in Apache Kudu. I know that we can query in Apache kudu using Apache Impala but i want to create some indexes in the Apache kudu to make the query processing faster,and my question is does Apache Kudu and Apache Impala support CREATE INDEX query and also what is the difference between partition and index.if i partition the Kudu table ,does that suffice for indexing ?

tk421
  • 5,775
  • 6
  • 23
  • 34
HJSG
  • 41
  • 1
  • 6

1 Answers1

2

Kudu has primary key concept, but it is the only index that it can support. Unlike any other RDBMS, there is no support for secondary index. But the primary key can be a composite key(Combination of multiple columns).

Since Indexing is only through primary key, I will try to explain the relation between primary key and partitioning. Kudu has several partitions called as Tablets which are located across multiple Tablet Servers. Enabling partitioning based on a primary key design will help in evenly spreading data across tablets. There are several partitioning techniques to achieve this, use case whether heavy read or heavy write will dictate the primary key design and type of partitioning.

Impala being a In-memory engine will make kudu much faster. There is a little advantage when compared to using impala with hive.

nanic
  • 76
  • 4