Awesome

hive-benchmarks

some benchmarking queries for Apache Hive

Setup

This repo was prepared for benchmarks of SS-DB, TPC-H and TPC-DS running in the following environment.

• Hadoop version: Hadoop 1.2.1
• Hive version: Hive 0.13-SNAPSHOT (Nov. 28, 2013)
• Cluster setup:
  • A 11-node (1 master + 10 slaves) EC2 cluster in us-east-1d
  • Instance type: m1.xlarge
  • OS Image: ami-a73264ce (Ubuntu Server 12.04.3 LTS 64-bit)
  • OS kernel image version: the result of cat /proc/version is Linux version 3.2.0-56-virtual (buildd@roseapple) (gcc version 4.6.3 (Ubuntu/Linaro 4.6.3-1ubuntu5) ) #86-Ubuntu SMP Wed Oct 23 09:43:22 UTC 2013.

Notes

Data types

Right now, int is used for the type of identifier. If the scale factor is very large, bitint is needed.

Because we may need to to compare the current version of Hive with a older version (e.g. 0.10.0) of it, we have to use data types supported by the older version to create columns. Here are mappings:

• decimal -> float
• char -> string
• vacahr -> string
• date -> string