Prashant Mishra
Developer Programs Engineer
Company: Google
Stream: В
Time: 16:00 - 16:45
Country: India
Language: English
Talk: Faster Select queries using open source java-bigquery
About the Speaker
Prashant is working as a Developer Programs Engineer with Google Cloud and he is responsible for Developing and Maintaining the open source BigQuery Java client library.Previously Prashant has worked as a Technical Solutions Consultant with Google where he was responsible for working on implementation of PWA, Chrome APIs and AMP technologies with Google’s top partners Before Google Prashant has worked as a Tech Lead with Cisco Systems mainly in the area of backend development
Talk: Faster Select queries using open source java-bigquery
The open source BigQuery Java client library makes it easier to access BigQuery APIs using Java. Since its inception in 2016, numerous features and improvements have been added to the client library to enable BigQuery developers and practitioners.
In this session we will talk about a new Beta method called `executeSelect` which we recently introduced in the new Connection interface of the Java client library. It supports read-only SELECT queries and provides many-folds faster query performance on high throughput workloads (reading 100 million rows).
We will talk about how this method is different from the existing ways of querying data from BigQuery with the help of a few code samples. We will also show how easy is it to integrate this method with your application if you are already familiar with JDBC, as this new method returns nothing but the familiar `java.sql.ResultSet`
Proposed Primary Take-away for the Audience
● Understand how to use this new performant method `executeSelect`
● Understand the difference between `executeSelect` and the current existing methods/APIs in terms of the way both has been implemented (Eg: REST vs GRPC being used under the hood)
Table of Contents:
● What is executeSelect
● How it differs from the existing APIs
● Discussion on the binary serialization and transport protocols which we are using
● Code samples showing how to use `executeSelect`
● Discussing on Performance improvements
● Where to submit feedback & what's next