A common design paradigm for a Data Lake is to ingest data into a “raw” zone with as little schema definition as possible. Often this involves landing the data in a semi-structured format, preferring a schema-on-read approach versus a schema-on-write. But when data scientists need to access this semi-structured data, it typically must be passed through a number of parsing and data prep jobs to make the data usable. In this session, I’ll demonstrate how you can use the features of Snowflake to implement the design pattern of schema-on-read, ingest semi-structured data into your Data Lake using a variety of techniques, and query your raw Data Lake zone with no parsing or data prep.
Bio: Kevin has spent over 20 years building data warehouses and data lakes at large enterprise companies on both the consulting side (Accenture, boutique) and the industry side. Prior to joining Snowflake, he owned a services firm that implemented Snowflake and he guided those customers in building new data architectures to make the most of their Snowflake investment. As a former Oracle ACE Director, Kevin is well equipped to help Oracle and other legacy data warehouse platform customers move to the cloud and adopt their workflows to take advantage of unique Snowflake capabilities