Subject: Re: Top 3 values for each group in PGSQL
To: pgsql-sql@lists.postgresql.org
References: <67FB61BF-CF8F-440E-B4C5-B6F72D61A9C7@gmail.com>
From: Thomas Kellerer <spam_eater@gmx.net>
Message-ID: <50d9b39f-56f8-8bc8-a93a-d5feb16f4dd3@gmx.net>
Date: Fri, 1 Mar 2019 13:52:02 +0100
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.1.23)
 Gecko/20090812 Thunderbird/2.0.0.23 Mnenhy/0.7.6.666
MIME-Version: 1.0
In-Reply-To: <67FB61BF-CF8F-440E-B4C5-B6F72D61A9C7@gmail.com>
Content-Type: text/plain; charset=utf-8
Content-Language: de-DE
Content-Transfer-Encoding: 8bit
Precedence: bulk

Ila B. schrieb am 01.03.2019 um 11:51:
> Hello,
> 
> I’m working on a health database and I’m trying to extract the most popular prescription codes from a custom table I structured like this:
> 
> Year - Code - Count(code)
> 
> I want to extract the 3 codes with maximum count for each year. I know I should be using rank() but I don’t really understand how this works.
> I am using pgAdmin4 version 3.5 with PostgreSQL 10.6 on Windows 10 Pro and no permission to update.

Something along the lines:

select code, year, "count"
from (
   select code, year, "count", 
          dense_rank() over (partition by code, year order by "count" desc) as rnk
   from the_table
) t
where rnk <= 3;