Subject: Re: [HACKERS] Surjective functional indexes
To: Alvaro Herrera <alvherre@alvh.no-ip.org>
Cc: Stephen Frost <sfrost@snowman.net>,
 Michael Paquier <michael.paquier@gmail.com>,
 Simon Riggs <simon@2ndquadrant.com>, Christoph Berg <myon@debian.org>,
 Thomas Munro <thomas.munro@enterprisedb.com>,
 Peter Eisentraut <peter.eisentraut@2ndquadrant.com>,
 PostgreSQL Hackers <pgsql-hackers@postgresql.org>
References: <20180323154534.4c3s3kdjil6hktyp@alvherre.pgsql>
From: Konstantin Knizhnik <k.knizhnik@postgrespro.ru>
Message-ID: <abaef102-e7a2-f86a-6df9-c6d92b65d989@postgrespro.ru>
Date: Fri, 23 Mar 2018 19:17:03 +0300
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101
 Thunderbird/52.4.0
MIME-Version: 1.0
In-Reply-To: <20180323154534.4c3s3kdjil6hktyp@alvherre.pgsql>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 7bit
Content-Language: en-US
Precedence: bulk


On 23.03.2018 18:45, Alvaro Herrera wrote:
> Konstantin Knizhnik wrote:
>
>> rd_projidx is not a list, it is Bitmapset. It is just one of many bitmap
>> sets in RelationData:
> Yes, but the other bitmapsets are of AttrNumber of the involved column.
> They new one is of list_nth() counters for items in the index list.
> That seems weird and it scares me -- do we use that coding pattern
> elsewhere?  Maybe use an array of OIDs instead?
>
Using list or array instead of bitmap requires much more space...
And bitmaps are much more efficient for many operations: check for 
element presence, combine, intersect,...
Check in bitmap has O(0) complexity so iterating through list of all 
indexes or attributes with bitmap check doesn't add any essential overhead.

Sorry, I do not understand this: "They new one is of list_nth() counters 
for items in the index list"

-- 
Konstantin Knizhnik
Postgres Professional: http://www.postgrespro.com
The Russian Postgres Company