Diving Into Delta Lake: Schema Enforcement & Evolution
Data, like our experiences, is always evolving and accumulating. To keep up, our mental models of the world must adapt to new data, some of which contains new dimensions - new ways of seeing things we had no conception of before. These mental models are not unlike a table's schema, defining how we categorize and...
Diving Into Delta Lake: Unpacking The Transaction Log
The transaction log is key to understanding Delta Lake because it is the common thread that runs through many of its most important features, including ACID transactions, scalable metadata handling, time travel, and more. In this article, we’ll explore what the Delta Lake transaction log is, how it works at the file level, and how...
Introducing Delta Time Travel for Large Scale Data Lakes
Data versioning for reproducing experiments, rolling back, and auditing data We are thrilled to introduce time travel capabilities in Databricks Delta, the next-gen unified analytics engine built on top of Apache Spark, for all of our users. With this new feature, Delta automatically versions the big data that you store in your data lake, and...
Benchmarking Structured Streaming on Databricks Runtime Against State-of-the-Art Streaming Systems
Update Dec 14, 2017: As a result of a fix in the toolkit’s data generator, Apache Flink's performance on a cluster of 10 nodes, multiple core cluster went from 6x slower than Apache Spark to 3x. This configuration is not examined in Apache Flink's post. In most uses cases, you will want to run streaming...
Running Streaming Jobs Once a Day For 10x Cost Savings
This is the sixth post in a multi-part series about how you can perform complex streaming analytics using Apache Spark. Traditionally, when people think about streaming, terms such as “real-time,” “24/7,” or “always on” come to mind. You may have cases where data only arrives at fixed intervals. That is, data appears every hour or...
Working with Complex Data Formats with Structured Streaming in Apache Spark 2.1
In part 1 of this series on Structured Streaming blog posts, we demonstrated how easy it is to write an end-to-end streaming ETL pipeline using Structured Streaming that converts JSON CloudTrail logs into a Parquet table. The blog highlighted that one of the major challenges in building such pipelines is to read and transform data...
New Features in Machine Learning Pipelines in Apache Spark 1.4
Apache Spark 1.2 introduced Machine Learning (ML) Pipelines to facilitate the creation, tuning, and inspection of practical ML workflows. Spark’s latest release, Spark 1.4, significantly extends the ML library. In this post, we highlight several new features in the ML Pipelines API, including: A stable API — Pipelines have graduated from Alpha! New feature transformers...
Using 3rd Party Libraries in Databricks: Apache Spark Packages and Maven Libraries
In an earlier post, we described how you can easily integrate your favorite IDE with Databricks to speed up your application development. In this post, we will show you how to import 3rd party libraries, specifically Apache Spark packages, into Databricks by providing Maven coordinates. Background on Spark Packages Spark Packages (http://spark-packages.org) is a community...
Making Databricks Better for Developers: IDE Integration
We have been working hard at Databricks to make our product more user-friendly for developers. Recently, we have added two new features that will allow developers easily use external libraries - both their own and 3rd party packages - in Databricks. We will showcase these features in a two-part series. Here is part 1, introducing how...
Statistical and Mathematical Functions with DataFrames in Apache Spark
We introduced DataFrames in Apache Spark 1.3 to make Apache Spark much easier to use. Inspired by data frames in R and Python, DataFrames in Spark expose an API that’s similar to the single-node data tools that data scientists are already familiar with. Statistics is an important part of everyday data science. We are happy...