question Re: Which is better for performance? hive 3 or hbase? in Support Questions

Which is better for performance? hive 3 or hbase?

Vinit_Kratin — Tue, 03 Sep 2019 14:46:57 GMT

I want to save events through Kafka into Hadoop ecosystem and also retrieve them with low latency. I can use hive3 or hbase. Which one should I use?

Re: Which is better for performance? hive 3 or hbase?

EricL — Tue, 03 Sep 2019 13:21:43 GMT

@Vinit_Kratin,

You can't compare Hive with HBase directly, it is really use case dependent. It depends on how you want to store you data, how you will query your data, and whether you need to delete data etc and etc.

You will need to explain a bit more on what you want to achieve before an answer can be given as to which is better, hive or hbase.

Doing a quick google search returned me below useful links:
https://community.cloudera.com/t5/Support-Questions/When-to-use-Hive-and-Hbase/m-p/206167
https://community.cloudera.com/t5/Community-Articles/The-Differences-between-Pig-Hive-and-HBase/ta-p/31007

https://www.dezyre.com/article/hive-vs-hbase-different-technologies-that-work-better-together/322

Cheers
Eric

Re: Which is better for performance? hive 3 or hbase?

Vinit_Kratin — Tue, 03 Sep 2019 14:10:40 GMT

Thanks a lot for your reply.
What I want to do is store events in the database.
These events are structured and needed to be queried multiple times.
Deletion is not required. low response time is a must.
Please let me know if I need to give you any more information.

Re: Which is better for performance? hive 3 or hbase?

Vinit_Kratin — Wed, 04 Sep 2019 06:21:42 GMT

@EricL

Re: Which is better for performance? hive 3 or hbase?

EricL — Wed, 04 Sep 2019 08:44:44 GMT

@ Vinit_Kratin,

If low response time is a must, then Hive is not an option, Impala should be faster.

HBase designed tables require a row key, which means, all data you want to retrieve will need this key to be queried against HBase. So for your event based data, if you need to use an event id to store and retrieve data, then hbase might be suitable.

I still suggest you to go through details of differences between HBase and Hive and Impala to get a better understanding. Google is your friend.

Cheers
Eric