Explore the intricacies of reading Apache Parquet files from the perspective of Apache Impala in this 26-minute conference talk. Delve into the crucial early stages of query execution in Apache Impala, focusing on the process from reading bytes of Parquet files on the filesystem to applying predicates and runtime filters on individual rows. Learn about Apache Impala's distributed massively parallel analytic query engine, optimized for both object stores and on-premises distributed file systems. Discover why Impala uses its own C++ Parquet scanner instead of existing libraries, enabling features like data caching, execution within memory bounds, and efficient parallelism. Gain insights into how these features give Impala an edge in the world of Big Data query engines. Presented by Csaba Ringhofer and Daniel Becker, experienced software engineers from Cloudera and members of the Apache Impala PMC, this talk offers valuable knowledge for those working with big data systems and file formats.
Overview
Syllabus
Anatomy of reading Apache Parquet files (from the Apache Impala perspective)
Taught by
The ASF