Tag: test

Waiting for PostgreSQL 19 – Add date and timestamp variants of random(min, max).

On 9th of September 2025, Dean Rasheed committed patch:

Add date and timestamp variants of random(min, max).
 
This adds 3 new variants of the random() function:
 
    random(min date, max date) returns date
    random(min timestamp, max timestamp) returns timestamp
    random(min timestamptz, max timestamptz) returns timestamptz
 
Each returns a random value x in the range min <= x <= max.
 
Author: Damien Clochard <damien@dalibo.info>
Reviewed-by: Greg Sabino Mullane <htamfids@gmail.com>
Reviewed-by: Dean Rasheed <dean.a.rasheed@gmail.com>
Reviewed-by: Vik Fearing <vik@postgresfriends.org>
Reviewed-by: Chao Li <li.evan.chao@gmail.com>
Discussion: https://postgr.es/m/f524d8cab5914613d9e624d9ce177d3d@dalibo.info

Continue reading Waiting for PostgreSQL 19 – Add date and timestamp variants of random(min, max).

How to play with upcoming, unreleased, PostgreSQL?

Lately reader of my blog asked about some detail, and then in next comment said that he doesn't have PostgreSQL 12 (which is currently developed version of Pg) available.

Well, I answered the original question, but I figured that if someone is reading my Waiting for … series, then it would make sense that such person could play with devel Pg herself.

So, here is how to get it done.

Continue reading How to play with upcoming, unreleased, PostgreSQL?

Should you use HASH index?

Today, Mattias|farm on IRC asked how to create primary key using HASH index. After some talk, he said that in some books it said that for “=" (equality) hash indexes are better.

So, I digged a bit deeper.

IMPORTANT UPDATE: As of PostgreSQL 10 hash indexes are WAL logged. As such, main point against them is gone.

Continue reading Should you use HASH index?

Test driven development for PostgreSQL

I have a mixed love/hate relationship with tests.

I hate writing them. I hate remembering to add them when I'm in the zone, and application code is flowing freely from the tips of my fingers.

But when I do add them, I absolutely love the ability to twist and replace the most core innards of application, and be able to tell that at least the sanity check of the code passes.

I love them even more when they prevent me for repeating some mistake/bug – i.e. when there is a bug, and I have tests (which clearly failed, as they didn't catch the bug), I add test for this specific bug, so I know that in future it will not happen again.

For a long time I've been fan of stored procedures (or functions) – of course not for everything, but where it makes sense.

Continue reading Test driven development for PostgreSQL