Just yesterday I committed that.

It is planned in PostgreSQL 9.2 to have support for cascading replication, which is the ability to add slaves under other slaves. In 9.2, slave-slave connections are only asynchronous.

After long weeks of battle, this week this commit has happened in Postgres-XC’s GIT repository.

Arrays can be created easily in PostgreSQL using the additional syntax [] when defining a column of a table.

This study is made with PostgreSQL 9.1.1, released a couple of days before this post is written. Unlogged tables are a new performance feature of PostgreSQL 9.1, created by Robert Hass. So, by guessing from this feature name, those tables are not logged in the database system :). More precisely, those tables do not use at all WAL (Write ahead log) that insure a safe database crash. Those tables are a good performance gain to contain data that do not especially need to survive from a crash, they are truncated automatically after a crash or unclean shutdown. Unlogged tables are shared among sessions, and are not deleted when a session ends. Autovacuum runs on them.

Scanning a table in a database can cost a lot in terms of CPU or I/O when calculating statistics in an application. The bigger application tables get, the more resource is necessary to calculate simple statistics for an application. A simple example of that are queries of the type:

In library git-core, git has a command that makes all the commits appearing in graphs. For example:

This post presents some basics when using foreign data wrappers with PostgreSQL for external files. FOREIGN DATA WRAPPER is a part of SQL/MED (Management of external data with SQL) and its implementation has begun since Postgres 8.4. This mechanism is based on COPY FROM to import data files directly into your database. Those tests have been done with 9.2 (development version).

The post presents a proposal to implement an HA solution based on PostgreSQL streaming replication and Standby node structure. This solution is still in construction, so the final implementation design that will be chosen for Postgres-XC may slightly change.

19th and 20th of May have been days of the PostgreSQL conference. During those 2 days, I saw some good presentations and stuff like that may help (perhaps) in increasing my own database knowledge related to PostgreSQL. By the way, for sure, what I heard from this conference will help me not only for my current work but also it could give ideas for future design tasks.