#334 — December 11, 2020

Read on the Web

Database Weekly

The Feature Casualties of Large Databases — We get it. You’ve got a lot of data, you’ve got to move and iterate fast, fires keep popping up and need to be put out and.. a lot of best practices and well intentioned designs fly out of the window. This post reflects on the most common ‘casualties’.

Brandur Leach

Comparing Fauna and DynamoDB — When the creator of a database compares their baby to someone else’s, you have to maintain a critical stance, but this is quite balanced and informative, showing where each shines, rather than screaming “Fauna rules!”

Evan Weaver (Fauna)

🧐 All Things Time-Series Modeling and Analysis — Get a primer on what time-series data is, when and why you need a time-series database, and how to use SQL to mine your data, answer questions, and predict trends ✨

Timescale sponsor

A Seamless MongoDB to PostgreSQL Migration — Running a digital currency exchange means having a solid foundation for your data systems is a must. Here Coinbase shares some lessons learned from a large cross-database data migration on a critical production system to AWS RDS PostgreSQL.

Alex Ghise (Coinbase)

Oracle Database 21c Bridges The NoSQL Gap — If you want it on prem you’ll be waiting till next year, but Oracle Database 21c is available in the cloud, at least, and the big new feature is heavily improved JSON support by way of a native JSON type.

The Register

TerminusDB Dropping GPLv3 for the Apache 2.0 License — Database system licensing changes have been more common in the past few years as vendors juggle wanting a healthy developer ecosystem against the threat of cloud providers reselling their software (as with MongoDB). The folks behind TerminusDB, an open source graph database, are moving from GPLv3 to Apache 2 to encourage independent vendors to embed the system into their apps.


👉 In related news, TerminusDB 4.0 has been released.

Amazon Redshift Announces Support for Native JSON and Semi-Structured Data Processing — AWS has added preview level support for JSON and semi-structured data to its Redshift managed data warehouse service, along with PartiQL support to query it.

Amazon Web Services

Kafka is Not a Database — Kafka is a message broker / streaming platform commonly used for processing streaming data and while it fills the role of a database in some specific ways, it’s not ideal as a primary data store, as demonstrated here.

Arjun Narayan and George Fraser

What You Need to Know About Aurora Serverless v2, So Far — Yan Cui asks (and answers!) a handful of questions you might be having about what Aurora Serverless v2 really means versus “v1.”

Yan Cui

How We Built Scalable Spatial Indexing in CockroachDB
Sumeer Bhola (Cockroach Labs)

🔨 Code and Tools

GNORM: Converts Database Schemas into Code or Text via Templates — Gnorm believes code shouldn’t generate database schemas, but database schemas should generate code!

Nate Finch et al.

query-exporter: Export Prometheus Metrics From SQL Queries — Supports PostgreSQL, MySQL, Oracle and Microsoft SQL Server via Python’s SQLAlchemy.

Alberto Donato

Announcing Prisma Migrate Preview - Simplified Database Migrations — Read the blog post on the Preview version of Prisma's newest release: Migrate, a data modeling and migrations tool.

Prisma sponsor

PoloDB: An Embedded JSON-Based Database — Always a fan to see these simple new databases built in languages like Go or Rust (as here). Has a Mongo-esque API but stores data locally a la SQLite.


OmniParser: A Go-Powered ETL System — Ingest CSV, text, JSON, and others, in a streaming fashion, and transform it into JSON based upon a schema. There’s an online demo here showing off some use cases.

JF Technology

🐦 Seen on Twitter

It's certainly not an unpopular opinion here at DB Weekly ;-) We 💙 SQL.