Neues vom PostgreSQL Planet

Heisenberg’s uncertainty principle

In quantum mechanics, the uncertainty principle […] asserts a fundamental limit to the precision with which the values for certain pairs of physical quantities of a particle, such as position, x, and momentum, p, can be predicted from initial conditions.

Uncertainty principle, Wikipedia

The functionality of using table partitions to speed up queries and make tables more manageable as data amounts grow has been available in Postgres for a long time already, with nicer declarative support available from v10 – so in general it’s a known technique for developers. But what is not so uniformly clear is the way how low-level partition management is done…as Postgres leaves it to users but no real standard tools or even concepts have emerged.

Event triggers are a very powerful mechanism to react to data structure changes in PostgreSQL.

Postgres already documents the ability to backup the database cluster using file system snapshots. Unfortunately, database clusters that use tablespaces often cannot use this method if the storage system doesn't support simultaneous snapshots across file systems.

PostgreSQL Person of the Week Interview with Paul Ramsey: I’m a middle-aged guy from the blessed west coast of Canada, where the sky and the mountains meet the sea. Raising a family takes up most of my time, but I’ve managed to carve out personal time for fitness recently, and exploring the area on my bike. I grew up in the interior of British Columbia, but have lived in Victoria, on Vancouver Island, for most of my adult life.

pgBackRest is a well-known powerful backup and restore tool. Old backups and archives are removed by the expire command based upon the defined retention policy.

Since the latest version published last month, new features regarding retention have been committed. We’ll here first overview those changes and then make a tour of the retention policy options that should be available in the next release.

Currently, it is impossible to move tables, indexes, and entire tablespaces from one cluster to another — that is because each table and index file is bound to the cluster's infrastructure because of:

If you want to see (and analyze) only a window of data over some continuous data stream in PostgreSQL, one way is to use a specialized tool like the PipelineDB extension. But if you can't do that, e.g. because you are stuck with AWS RDS or for some other reason, streaming data tables, or continuous views, can be implemented with pretty much PostgreSQL alone.

Missing Piece: Failover of the Logical Replication Slot

Logical decoding capability has existed in PostgreSQL for the last several versions, and a lot more functionalities are added over time.  However, one of the very crucial functionalities is still missing in PostgreSQL, which prevents logical replication to be part of critical production environments where downstream expects the logical changes to be streamed over a database connection reliably – even after a high-availability failover.

PostgreSQL 13 is in beta right now, and we need everyone to test it! The RPMs are out. Continue reading "Installing PostgreSQL 13 beta/RC on RHEL/CentOS/Fedora"

This article gives a step by step guide to utilizing Machine Learning capabilities with 2UDA. In this article, we will use examples of Animals to predict whether they are Mammals, Birds, Fish or Insects. Software versions We are going to use 2UDA version 11.6-1 to implement the Machine Learning model. This version 11.6-1 combines: PostgreSQL […]

This year, I was so excited about doing a workshop about optimizing Python & Django apps with Postgres superpowers for the PyCon 2020 conference.

Working with other developers on performance is something I always find amazing. So props to the Python people at Microsoft who encouraged my team to create a workshop on Postgres for PyCon 2020. Thank you to Nina Zakharenko, Dan Taylor, & Crystal Kelch.

CREATE TABLE one_column_table (field INTEGER);

The above SQL creates ‘one_column_table’ with only a single column. But does the table actually have only one column?

Let’s query the pg_attribute catalog to find out how many columns our one_column_table has.

This email thread is illustrative of why it is unwise to place the Postgres data directory (PGDATA) at the top of a mount point. Instead, create a subdirectory under the mount point and put PGDATA there. This has the advantage of avoiding possible data corruption if mounting fails, and allows more efficient use of pg_upgrade.

MQTT has become a de-facto standard for the transport of messages between IoT devices. As a result, a plethora of libraries and MQTT message brokers have become available. Can we use this to transport messages originating from PostgreSQL?

Recently we have covered “count” quite extensively on this blog. We discussed optimizing count(*) and also talked about “max(id) – min(id)” which is of course a bad idea to count data in any relational database (not just in PostgreSQL). Today I want to focus your attention on a different kind of problem and its solution: Suppose you want to grant a user access to a certain piece of data only X times. How can one implement that safely?

I am planning to virtually attend and present at the Percona Live Online conference tomorrow, May 19. It starts at 10am, Eastern USA time, and spans 24 hours, so it covers every time zone. I am speaking at noon, Eastern USA time.

Attendance is free, so you might want to check it out. I saw some interesting topics on the program. I am also curious to experience a 24-hour virtual conference, though I am unlikely to remain awake that long.

pgBackRest is a well-known powerful backup and restore tool. The 2.26 version has been released on Apr 20, 2020. New features have been developed since then.

Today, let’s have a look at: add backup/expire running status to the info command.

 

We are happy to announce the new release of PgMex 1.2.0!