Therefore, BlinkDB supports two types of joins:ġ) Arbitrary joins are supported if the join key is present in the columns set of one of the stratifies samples. For example, instead of automatically creating samples for a dataset, this version supports manual sample creation with explicitly specified sample ratio only.īlinkDB does not build online samples, so it does not build samples for the joined tables. The latest version number is alpha-0.2.0. It has two major parts: one is the sample building engine that selects what stratified samples to build by considering historic workloads and the distribution of the data the other part is a dynamic sample selection module that chooses appropriate sample files at runtime according to specific time/accuracy requirements specified by the query.īlinkDB has a public open source repository on Github.
The number and sizes of the stratified samples are limited by the storage budget specified when importing the data. BlinkDB builds a couple of stratified samples on the original data and executes the queries on the samples instead of the original data to reduce query execution time. It allows users to trade-off query accuracy for response time, thus enabling interactive queries on big data.
BlinkDB is an approximate query engine built on top of Hive and Shark ("Hive on Spark", the former Spark SQL).
0 kommentar(er)
