Скачать с ютуб A deep dive into the TileDB data format & storage engine в хорошем качестве

A deep dive into the TileDB data format & storage engine

For a summary and links to the Jupyter notebooks used in this presentation, see the blog post at https://tiledb.com/blog/a-deep-dive-i... . Everything in TileDB is built around the open-source array storage engine, TileDB Embedded. Any data can be universally modeled as multi-dimensional TileDB arrays, and it all starts here. Complete with code examples and performance tips, this presentation provides a thorough technical primer for getting started with TileDB Embedded. It also lays the foundation for getting started with the universal database, TileDB Cloud. You will learn: • Why we chose arrays as the foundational data structure of TileDB • How dataframes are just a special instance of arrays • Data layout, tiling, compression and other filters • Dense and sparse array indexing for rapid slicing • Data versioning and time traveling • Why TileDB is “cloud-optimized” • How to achieve unprecedented performance • Going beyond storage: compute push-down • How to evolve and manage schemas in TileDB Contents of this video 0:00:00 – Introduction & Why arrays 0:13:15 – The basics 0:32:46 – Tiling & layout 0:42:01 – Advanced internal mechanics 0:59:18 – Performance tips 1:02:43 – Work in progress 1:04:37 – TileDB vs. others 1:08:06 – Q&A About TileDB makes data management and compute fast, easy and universal. Manage any data as multi-dimensional arrays and access with any tool at global scale. Connect with us Website: https://tiledb.com/ Twitter:   / tiledb   LinkedIn:   / tiledb-inc   Book a personalized product demo: https://tiledb.com/demo Sign up at https://cloud.tiledb.com/auth/signup and contact [email protected] for free credits.