Обсуждение: Which file does the SELECT?

On sön, 2010-10-10 at 13:32 +0530, Vaibhav Kaushal wrote:
> I have gone through the source code a bit but I wanted to know that
> which file contains the code that performs the final SLECTION after
> the optimizer has created the final plan? I mean which part of the
> executor is responsible for the SELCT to be run? 

That depends on what plan was chosen for the SELECT, since the executor
is primarily organized by plan node type, independent of which statement
caused the plan to be generated.

2010/10/10 Vaibhav Kaushal <vaibhavkaushal123@gmail.com>:
> Thanks for the reply.
> So if I am not wrong, I will have to understand the whole querying process
> in detail? If it is so, then where do I start from?
> -Vaibhav
>
> On Sun, Oct 10, 2010 at 1:41 PM, Peter Eisentraut <peter_e@gmx.net> wrote:
>>
>> On sön, 2010-10-10 at 13:32 +0530, Vaibhav Kaushal wrote:
>> > I have gone through the source code a bit but I wanted to know that
>> > which file contains the code that performs the final SLECTION after
>> > the optimizer has created the final plan? I mean which part of the
>> > executor is responsible for the SELCT to be run?
>>
>> That depends on what plan was chosen for the SELECT, since the executor
>> is primarily organized by plan node type, independent of which statement
>> caused the plan to be generated.
>>
> So if I am not wrong, I will have to understand the whole querying process
> in detail? If it is so, then where do I start from?
>

And it depends on what you are interested in. If the executor behavior
is your interest, see execMain.c, but your question looks more
interested in nodeSeqscan.c which scans rows from a relation.

Regards,



--
Hitoshi Harada

Vaibhav Kaushal <vaibhavkaushal123@gmail.com> writes:
> So if I am not wrong, I will have to understand the whole querying process
> in detail? If it is so, then where do I start from?

If you haven't seen it already, this is a good place to start:
http://developer.postgresql.org/pgdocs/postgres/overview.html

There's also some stuff in src/tools/backend/ that tries to summarize
which subdirectories of the source tree do what.

Also, many subsystems have README files in the source tree.
It sounds like you might like to look at src/backend/executor/README.

But in the end there's no substitute for reading source code...
        regards, tom lane

Vaibhav Kaushal <vaibhavkaushal123@gmail.com> writes:
> Can someone tell me what are 'Join Pairs with no Join clause' ? I am not
> able to figure that out!

Consider
select * from t1, t2, t3 where t1.a = t2.x and t1.b = t3.y;

In theory this query could be done by first joining t2 and t3, then
joining that to t1.  But the planner won't investigate the possibility
because the t2/t3 join would have to be a cartesian product join:
there's no WHERE clause relating them.

On the other hand, if we have
select * from t1, t2, t3 where t1.a = t2.x and t1.a = t3.y;

then the planner is able to infer the additional join clause t2.x =
t3.y, so it will consider that join sequence.
        regards, tom lane

On Sun, Oct 10, 2010 at 10:51:54PM +0530, Vaibhav Kaushal wrote:
> However, I find too many references to the Data structure "datum" what is it
> and where is it defined? Can someone tell me please? Also, what role does it
> play?

"Datum" is the singular form of "data". It refers to a single item of
any type. So it may be an integer, text type, geometry type, anything.
A row is a list of datums. (A datum can also be a composite type).

Normally in the planner code you don't need to worry to much about what
it explicitly refers to, but if you do, you need to know the type of a
datum before you can manipulate it. The type is not stored inside the
datum.

Have a nice day,
--
Martijn van Oosterhout   <kleptog@svana.org>   http://svana.org/kleptog/
> Patriotism is when love of your own people comes first; nationalism,
> when hate for people other than your own comes first.
>                                       - Charles de Gaulle

On Oct 10, 2010, at 12:21 PM, Vaibhav Kaushal wrote:

> Thanks to both hitoshi and tom for your replies.
>
> I think I need to look into the Postgres code itself (I am better at code than documentation). But since I have not
beentouch with C lately (these days I am programming on PHP) I think I have forgot a few rules of game (afterall PHP is
somuch more easy than C :P ). Moreover,  postgres is the first Open Source software whose code I am interested in. I
havenever looked into other OSS codes much except correcting a few compilation errors here and there on beta / alpha
releases. 
>
> I have had the chance and success to compile my own Linux OS and it was fun to do so... but I guess development is a
tougherjob. With an idea in mind, and a thankful feeling towards postgres is what drives me to do this tougher job. 
>
> When I was designing my database for a web app, I found so many problems in MySQL that I could not continue (the best
ofall, I can't use the commands written in my DB book to create a foreign key, it does not natively support foreign
keys,confusing storage engines and so on).. and then I got postgres which I am a fan of. 
>
> I hope I will not be flamed when I will ask those questions (some of them are actually very silly ones).
>
>  I will look inside the code now and will get back after i get some progress with it.
>
> However, I find too many references to the Data structure "datum" what is it and where is it defined? Can someone
tellme please? Also, what role does it play? 
>
> Thanks to you all for your replies.
>
> -Vaibhav

Depending on your text editor, you may be able to utilize TAGS files; see src/tools/make_(e|c)tags for creating TAGS
filesfor your editor of choice (emacs/vim, although other editors may support specific formats).  This will allow you
tonavigate to the specific definition of the type/function/macro, and can be very enlightening and help answer some of
thesequestions.  `git grep` will also come in handy if you're working directly from a git checkout. 

Regards,

David
--
David Christensen
End Point Corporation
david@endpoint.com

On Mon, Oct 11, 2010 at 04:14:04PM +0530, Vaibhav Kaushal wrote:
> Is there something like that in Kdevelop? I dont use CLI editors much.

KDevelop is listed as one of the editors that support ctags.

http://en.wikipedia.org/wiki/Ctags

Cheers,
David.
-- 
David Fetter <david@fetter.org> http://fetter.org/
Phone: +1 415 235 3778  AIM: dfetter666  Yahoo!: dfetter
Skype: davidfetter      XMPP: david.fetter@gmail.com
iCal: webcal://www.tripit.com/feed/ical/people/david74/tripit.ics

Remember to vote!
Consider donating to Postgres: http://www.postgresql.org/about/donate

David Fetter <david@fetter.org> writes:

> On Mon, Oct 11, 2010 at 04:14:04PM +0530, Vaibhav Kaushal wrote:
>> Is there something like that in Kdevelop? I dont use CLI editors much.
>
> KDevelop is listed as one of the editors that support ctags.

I've just developed some code for the backend and used cscope (from
Emacs directly) which is excellent.
 http://cscope.sourceforge.net/ http://kscope.sourceforge.net/ http://www.ziplink.net/~felaco/cbrowser/

Unfortunately the KDE GUI for it seems unmaintained now.

Regards,
-- 
Dimitri Fontaine
http://2ndQuadrant.fr     PostgreSQL : Expertise, Formation et Support

On Mon, Nov 8, 2010 at 9:55 PM, Vaibhav Kaushal
<vaibhavkaushal123@gmail.com> wrote:
> I have started with the work and am using Eclipse and it helps quite a lot.
> I can find the declarations quite easily. Thanks to open Source.
>
> BTW, I am encountering too many (just too many) data types as I try to
> understand the backend (specifically the executor). I do think that its
> normal because the executor has to consider almost everything that the other
> parts of the DB can possibly command it to do.
>
> Is there some documentation available on the data types / structures which
> are in use at the backend?

There's less than one might hope.  I think you pretty much have to
look through the README files and source code comments.

-- 
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

Yeah that is what seems to be the best way. The thing is that I am looking in the PostgreSQL code for the first time
andI am not fully aware of the data structures or the methods / algos implemented in the project. This makes the work
offinding out whats and whys much more difficult. So I asked it on the list.<br /><br />Thanks anyways for the reply.
<br/><br />-Vaibhav (*_*)<br /><br /><div class="gmail_quote">On Wed, Nov 10, 2010 at 9:38 AM, Robert Haas <span
dir="ltr"><<ahref="mailto:robertmhaas@gmail.com">robertmhaas@gmail.com</a>></span> wrote:<br /><blockquote
class="gmail_quote"style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left:
1ex;"><divclass="im">On Mon, Nov 8, 2010 at 9:55 PM, Vaibhav Kaushal<br /> <<a
href="mailto:vaibhavkaushal123@gmail.com">vaibhavkaushal123@gmail.com</a>>wrote:<br /> > I have started with the
workand am using Eclipse and it helps quite a lot.<br /> > I can find the declarations quite easily. Thanks to open
Source.<br/> ><br /> > BTW, I am encountering too many (just too many) data types as I try to<br /> >
understandthe backend (specifically the executor). I do think that its<br /> > normal because the executor has to
consideralmost everything that the other<br /> > parts of the DB can possibly command it to do.<br /> ><br />
>Is there some documentation available on the data types / structures which<br /> > are in use at the backend?<br
/><br/></div>There's less than one might hope.  I think you pretty much have to<br /> look through the README files and
sourcecode comments.<br /><font color="#888888"><br /> --<br /> Robert Haas<br /> EnterpriseDB: <a
href="http://www.enterprisedb.com"target="_blank">http://www.enterprisedb.com</a><br /> The Enterprise PostgreSQL
Company<br/></font></blockquote></div><br />