SELECT 'amazing features' FROM "postgresql" by Kevin Davin

Discover advanced PostgreSQL features and best practices with speaker Kevin Davin.

Key takeaways
  • Types in PostgreSQL: Kevin Davin introduces types into databases in real-time, which allows data to be validated during insertion.
  • Avoiding code smells: He stresses the importance of avoiding code smells, such as the use of NOT IN clauses, and suggests using foreign data wrappers instead.
  • PG L-tree: Davin explains how to use PostgreSQL’s L-tree data structure to efficiently query large amounts of data.
  • UUIDs: He recommends using universally unique identifiers (UUIDs) instead of row numbers or serial values to ensure data integrity.
  • Commentable expressions: Davin demonstrates how to use commentable expressions to create recursive queries in PostgreSQL.
  • Foreign data wrappers: He explains how foreign data wrappers can be used to integrate PostgreSQL with other data sources, such as file systems or other databases.
  • Window functions: Davin shows how window functions can be used to perform calculations over sets of rows that are related to the current row.
  • Data transformation: He emphasizes the importance of data transformation and validation when integrating data from different sources.
  • UUIDs as keys: Davin suggests using UUIDs as keys in PostgreSQL to ensure data integrity and avoid conflicts.
  • row numbers vs. UUIDs: He compares the use of row numbers to the use of UUIDs as primary keys and recommends using UUIDs for improved data integrity.
  • PostgreSQL’s strengths: Davin highlights the strengths of PostgreSQL, including its ability to handle complex queries, integration with other data sources, and support for advanced data types.
  • UUIDs in production: He notes that UUIDs have been used successfully in production environments and are a good choice for ensuring data integrity.
  • Best practices: Davin emphasizes the importance of following best practices when working with PostgreSQL, including using foreign data wrappers and UUIDs.