How does data flow to hive client

[#1]

Hello,

I have hive client installed on my Linux machine, If I issue a query , would output data of the query flow directly from data nodes to my machine or would it go to hive server ( it's installed on a node other than the data nodes ) and then arrive at my machine?

Thanks in advance.

[#2 · Accepted solution]

This might be a good read:

http://www.guru99.com/introduction-hive.html

[#3]

Depending on the client the data flow from the query works differently. In case of using Hive CLI, you access the HDFS directly to get the data, and Hive Metastore for table metadata. In case of HiveServer2 client (ex Beeline CLI) all data transmission happens through HiveServer2. HiveServer2 is recommended from security and scalability perspective.