Re: A SQL Training - Mailing list pgsql-sql

From Achilleus Mantzios
Subject Re: A SQL Training
Date
Msg-id Pine.LNX.4.44.0207170948400.9911-100000@matrix.gatewaynet.com
Whole thread Raw
In response to A SQL Training  (Devrim GUNDUZ <devrim@oper.metu.edu.tr>)
List pgsql-sql
On Tue, 16 Jul 2002, Devrim GUNDUZ wrote:

You need to identify your problem: preorder traversal on a tree
using pointers for the fathers.
Then you have to prove (or prove the opposite) that SQL alone
is capable of expressing what you need.

BTW, Nested Sets are only good for "static" databases.
Pointers are good for intensively "dynamic" databases, but have some
performance problems (for instance traversing to the root).
A genealogical approach (just like Oleg's and Teodor's one in 
contrib/tree) seems the most attractive.
Also you can have your own setup using postgresql arrays
(The arrays could contain the path of ids of the nodes from
the specific node to the root).
All these probably have little to do with your problem
(which is a computational theory problem), but maybe useful
in other cases.

> 
> Hi,
> 
> Today, one of my teacher asked me a question for a practice... I could not 
> solve it :)
> 
> Here it comes:
> 
> We are given the name of the grandfather in a family. Let's call him A. A 
> has n sons, each son has children .
> 
> for example.
>           A -> Grandfather
>              / \
>             /   \
>               B     C
>           /\   / \ \ 
>               /  \ D  E  F
>             G   H
>                 ...
> 
> We are given the data like ('A','B'),('C','D')...
> 
> Now, we do not know how many children are in the list. How could we list 
> this family tree in "tree" formt by only using SQL?
> 
> Best regards.
> 
> 
> 
> 
> 
> 

-- 
Achilleus Mantzios
S/W Engineer
IT dept
Dynacom Tankers Mngmt
Nikis 4, Glyfada
Athens 16610
Greece
tel:    +30-10-8981112
fax:    +30-10-8981877
email:  achill@matrix.gatewaynet.com       mantzios@softlab.ece.ntua.gr



pgsql-sql by date:

Previous
From: Rudi Starcevic
Date:
Subject: Re: Sequence name length
Next
From: "Rajesh Kumar Mallah."
Date:
Subject: Re: Cascading deletions does not seem to work inside PL/PGSQL functions.