Crunchy Data makes PostgreSQL lessons interactive

Learning database administration requires access to a database on which mistakes are affordable. A written explanation can describe a command, but an exercise lets the reader change an input and inspect the result. Crunchy Data's new learning portal puts interactive work around focused PostgreSQL and PostGIS subjects.

The December 26 announcement describes self-directed modules built on Katacoda. Each tutorial takes approximately twenty minutes. The intended readers include system administrators, database administrators and application developers, with subjects spanning basic administration, backup and restore with pgBackRest, JSON data types and PostGIS.

I like this approach because PostgreSQL exposes mechanisms worth learning directly. Query plans and schema behavior become easier to understand when a reader can inspect them. A bounded exercise can provide that opportunity without requiring the reader to assemble every part of an environment before encountering the subject.

A short module needs a concrete question

The announcement divides the portal into administration, application development and PostGIS modules. Administration material introduces basic operations and PostgreSQL capabilities. For developers, the portal covers introductory and more advanced application topics. PostGIS includes geometries, geography and spatial relationships.

That division reflects different questions a reader brings to the database. An administrator needs to understand how an operation affects the installation and its data. A developer needs to understand what a query or data type means. Spatial work introduces another set of concepts that ordinary table examples do not explain on their own.

Approximately twenty minutes is enough for a focused exercise, but it is not a substitute for complete operational preparation. The value is the opportunity to connect one explanation with an observable result. A module can establish that connection and leave the reader with a more precise question for the full documentation.

The announcement's inclusion of backup and restore is useful in this respect. Backup tooling is easier to understand when recovery is part of the subject from the beginning. A copy operation and a usable recovery procedure are related concerns, but completing the former does not by itself demonstrate the latter.

This is a judgment about the curriculum described in the announcement, not a claim that completing its lessons certifies an installation. The source promises focused interactive tutorials. It does not give an assessment standard for production readiness or claim that every administrative failure is covered.

Query plans show why interaction helps

PostgreSQL's own performance documentation shipped with 12.0 gives a concrete example of why changing inputs helps learning. It explains how progressively more selective conditions can lead the planner to choose different scans.

A sequential scan with a filter still visits the table's rows, even when the filter reduces the number emitted. The plan's estimated output count therefore does not equal the amount of scanning work. That distinction is easier to retain when two related queries show different output estimates while retaining a sequential scan.

The manual then narrows the condition enough for a bitmap index scan and heap scan to become worthwhile. The index identifies row locations, and the heap scan retrieves the rows in a way that reduces the cost of separate fetches. A still more selective equality condition can favor a simple index scan.

These are examples from the PostgreSQL manual, not a description of a particular portal lesson. They illustrate a useful criterion for interactive instruction. Changing one predicate can make an implementation choice visible, while the explanation supplies the reason for that choice.

The manual also warns that estimated costs and row counts can vary. Statistics come from samples, and costs use planner parameters rather than direct units of elapsed time. A useful exercise should help readers understand that variation instead of treating one printed plan as the only correct result.

I consider that a better goal than memorizing command output. PostgreSQL's planner provides estimates that can be inspected and questioned. Learning what a number represents is more transferable than recognizing the exact number from a tutorial.

Open lessons can be examined and adapted

The announcement makes the tutorial material available under Creative Commons Attribution-Share Alike 4.0. It also points readers to the project's GitHub repository and invites issues. That gives the learning material a defined reuse model and a public route for reporting problems.

The tutorial license should be distinguished from a promise about every part of the hosted learning service. The announcement describes the lessons as openly licensed and names Katacoda as the platform used to deliver them. Those are different components of the learning environment.

Reusable instructional material is particularly useful for database topics because examples need context. An explanation can accompany the commands that produce the result, rather than circulating as an isolated snippet. The shared material can retain its rationale when adapted for another instructional setting, subject to its license.

The portal's scope also recognizes PostgreSQL's extension ecosystem through a dedicated PostGIS module. Spatial relationships require their own explanation even when the underlying installation remains PostgreSQL. Giving them a focused course avoids assuming that familiarity with ordinary relational queries supplies all the necessary spatial concepts. The announced portal combines that subject-specific teaching with shorter administration and development exercises, each intended to let the reader work through a defined PostgreSQL topic interactively.