473,836 Members | 1,574 Online
Bytes | Software Development & Data Engineering Community
+ Post

Home Posts Topics Members FAQ

Query Planner Filtering Of Specified Value From 'On Distinct' Clause

I have a table from which I'm trying to extract certain information. For
historical reasons, we archive every action on a particular thing
('thing' is identified, funnily enough, by 'id'). So the only way to
find out the current state of a particular combination of attributes is
to "select distinct on (id, ...) ... order by date desc".

In the examples below, I've taken real output from psql and done a
global search/replace on various names to convince myself that I'm not
leaking information. The numbers are intact, just the column/table/view
names are different.

For sake of argument, here is the table structure:
database=> \d souce
Table "public.sou rce"
Column | Type | Modifiers
id | integer | not null
date | timestamp without time zone | default 'now'
second | integer | not null
third | text | not null
Indexes: source btree (id)
This table contains about 98000 rows. Let's create a view into the
"source" table:

database=> create view myview as select distinct
on (id, second, third) *
from source
order by id, second, third, date desc;

Now let's look at the query plan for "select * from myview":

database=> explain analyze select * from myview;
Subquery Scan myview (cost=12483.55. .13463.96 rows=9804 width=63)
(actual time=1262.08..1 450.68 rows=21089 loops=1)
-> Unique (cost=12483.55. .13463.96 rows=9804 width=63)
(actual time=1262.06..1 406.71 rows=21089 loops=1)
-> Sort (cost=12483.55. .12728.65 rows=98041 width=63)
(actual time=1262.06..1 339.34 rows=98063 loops=1)
Sort Key: source.id, source.second, source."third",
-> Seq Scan on source (cost=0.00..224 7.41 rows=98041
width=63) (actual time=0.02..269. 07 rows=98063 loops=1)
Total runtime: 1467.78 msec
(6 rows)

I'm not sure if there's really anything I can do to make that go faster.
But look what happens when I specify a particular value for one of the
fields in the "on distinct" clause:

database=> explain analyze select * from myview where id=12345;
Subquery Scan myview (cost=12483.55. .13463.96 rows=9804 width=63)
(actual time=1446.58..1 446.81 rows=2 loops=1)
Filter: (id = 12345)
-> Unique (cost=12483.55. .13463.96 rows=9804 width=63)
(actual time=1258.87..1 434.39 rows=21089 loops=1)
-> Sort (cost=12483.55. .12728.65 rows=98041 width=63)
(actual time=1258.87..1 336.89 rows=98063 loops=1)
Sort Key: source.id, source.second, source."third",
-> Seq Scan on source (cost=0.00..224 7.41 rows=98041 width=63)
(actual time=0.02..268. 54 rows=98063 loops=1)
Total runtime: 1457.53 msec
(7 rows)
Hmm.. not as efficient as I would have liked. The planner happily
executed the whole view, and only then filtered out just the bits that I
wanted. The most expensive parts of the query were done on the
unfiltered set. The result set is only 0.002% of the whole data set, so
I was a little shocked that it took *longer* to generate the results
that I wanted, than it took to create the unfiltered results!

Is there any reason why it would not be sensible to push the filter down
to just after the first sequential scan of source? Since "id" is one of
the fields in the "on distinct" condition, doesn't it make sense that
pushing the filter further down would reduce the amount of work done to
get exactly the same result? The same would be true if I was to filter
on a specific "second" or "third" value.

I understand that filtering on columns not present in the "on distinct"
clause would have drastically different effects when done before or
after the "Unique" processing. However, since the column(s) I'm
filtering on are only those present in the "on distinct" clause, it
makes no difference whether the filter is before or after the

Am I totally stark, raving mad?
Alex Satrapa
---------------------------(end of broadcast)---------------------------
TIP 6: Have you searched our list archives?


Nov 12 '05 #1
1 2492
Alex Satrapa <al**@lintelsys .com.au> writes:
Is there any reason why it would not be sensible to push the filter down
to just after the first sequential scan of source?

Try 7.4 ...

regards, tom lane

---------------------------(end of broadcast)---------------------------
TIP 5: Have you checked our extensive FAQ?


Nov 12 '05 #2

This thread has been closed and replies have been disabled. Please start a new discussion.

Similar topics

by: Jonathan | last post by:
Hi all! For a match schedule I would like to find all possible combinations of teams playing home and away (without teams playing to themselves of course). I now the simple version works something like this: For a (very) simple table containing three rows like this: row 1: A
by: Ed_No_Spam_Please_Weber | last post by:
Hello All & Thanks in advance for your help! Background: 1) tblT_Documents is the primary parent transaction table that has 10 fields and about 250,000 rows 2) There are 9 child tables with each having 3 fields each, their own PK; the FK back to the parent table; and the unique data for that table. There is a one to many relation between the parent and each of the 9 child rows. Each child table has between 100,000 and 300,000
by: Együd Csaba | last post by:
Hi All, how can I improve the query performance in the following situation: I have a big (4.5+ million rows) table. One query takes approx. 9 sec to finish resulting ~10000 rows. But if I run simultaneously 4 similar queries it takes nearly 5 minutes instead of 4 times 9 seconds or something near of that. here is a sample query: select mertido, fomeazon, ertektipus, mertertek from t_me30 where fomeazon in (select distinct fomeazon...
by: Riley DeWiley | last post by:
I am continually amazed by SQL's ability to humble me .... I have a toy query into a toy database that looks just like this: SELECT . FROM f, fw, w WHERE f.id = fw.fid and fw.wid = w.id and w.id in (((((select distinct w.id from w where w.x in ('a', 'b') ))))) GROUP BY . HAVING count(f.id) = (select count(id) from (((((select distinct w.id from w
by: CSN | last post by:
I have a pretty simple select query that joins a table (p) with 125K rows with another table (pc) with almost one million rows: select p.* from product_categories pc inner join products p on pc.product_id = p.id where pc.category_id = $category_id order by p.title
by: Zygo Blaxell | last post by:
I have a table with a few million rows of temperature data keyed by timestamp. I want to group these rows by timestamp intervals (e.g. every 32 seconds), compute aggregate functions on the columns, and ultimately feed the result into a graph-drawing web thingy. I'm trying a few different ways to get what seems to be the same data, and seeing some odd behavior from the query planner. The table looks like this:
by: Jeff Boes | last post by:
I hope this helps someone else ... I had struggled some time ago with attempts to get a rank of values query to work, but then I gave up and set it aside. I had another reason to attack it, and in between then and now I learned how to return "setof" values from a function, as well as how to construct "dynamic" queries inside a function. Returning the top 10 values from a query is no big deal: create table my_table (field1 integer,...
by: clare at snyder.on.ca | last post by:
I have a SQL query I need to design to select name and email addresses for policies that are due and not renewed in a given time period. The problem is, the database keeps the information for every renewal in the history of the policyholder. The information is in 2 tables, policy and customer, which share the custid data. The polno changes with every renewal Renewals in 2004 would be D, 2005 S, and 2006 L. polexpdates for a given customer...
by: chrisdavis | last post by:
I'm trying to filter by query or put those values in a distinct query in a where clause in some sort of list that it goes through but NOT at the same time. Example: ROW1 ROW2 ROW3 ROW4 , etc. I want to go to the first row, do a WHERE statement, return the
by: marktang | last post by:
ONU (Optical Network Unit) is one of the key components for providing high-speed Internet services. Its primary function is to act as an endpoint device located at the user's premises. However, people are often confused as to whether an ONU can Work As a Router. In this blog post, we’ll explore What is ONU, What Is Router, ONU & Router’s main usage, and What is the difference between ONU and Router. Let’s take a closer look ! Part I. Meaning of...
by: Hystou | last post by:
Most computers default to English, but sometimes we require a different language, especially when relocating. Forgot to request a specific language before your computer shipped? No problem! You can effortlessly switch the default language on Windows 10 without reinstalling. I'll walk you through it. First, let's disable language synchronization. With a Microsoft account, language settings sync across devices. To prevent any complications,...
by: Oralloy | last post by:
Hello folks, I am unable to find appropriate documentation on the type promotion of bit-fields when using the generalised comparison operator "<=>". The problem is that using the GNU compilers, it seems that the internal comparison operator "<=>" tries to promote arguments from unsigned to signed. This is as boiled down as I can make it. Here is my compilation command: g++-12 -std=c++20 -Wnarrowing bit_field.cpp Here is the code in...
by: Hystou | last post by:
Overview: Windows 11 and 10 have less user interface control over operating system update behaviour than previous versions of Windows. In Windows 11 and 10, there is no way to turn off the Windows Update option using the Control Panel or Settings app; it automatically checks for updates and installs any it finds, whether you like it or not. For most users, this new feature is actually very convenient. If you want to control the update process,...
by: tracyyun | last post by:
Dear forum friends, With the development of smart home technology, a variety of wireless communication protocols have appeared on the market, such as Zigbee, Z-Wave, Wi-Fi, Bluetooth, etc. Each protocol has its own unique characteristics and advantages, but as a user who is planning to build a smart home system, I am a bit confused by the choice of these technologies. I'm particularly interested in Zigbee because I've heard it does some...
by: agi2029 | last post by:
Let's talk about the concept of autonomous AI software engineers and no-code agents. These AIs are designed to manage the entire lifecycle of a software development project—planning, coding, testing, and deployment—without human intervention. Imagine an AI that can take a project description, break it down, write the code, debug it, and then launch it, all on its own.... Now, this would greatly impact the work of software developers. The idea...
by: isladogs | last post by:
The next Access Europe User Group meeting will be on Wednesday 1 May 2024 starting at 18:00 UK time (6PM UTC+1) and finishing by 19:30 (7.30PM). In this session, we are pleased to welcome a new presenter, Adolph Dupré who will be discussing some powerful techniques for using class modules. He will explain when you may want to use classes instead of User Defined Types (UDT). For example, to manage the data in unbound forms. Adolph will...
by: conductexam | last post by:
I have .net C# application in which I am extracting data from word file and save it in database particularly. To store word all data as it is I am converting the whole word file firstly in HTML and then checking html paragraph one by one. At the time of converting from word file to html my equations which are in the word document file was convert into image. Globals.ThisAddIn.Application.ActiveDocument.Select();...
by: adsilva | last post by:
A Windows Forms form does not have the event Unload, like VB6. What one acts like?

By using Bytes.com and it's services, you agree to our Privacy Policy and Terms of Use.

To disable or enable advertisements and analytics tracking please visit the manage ads & tracking page.