Eliminating Combinatorial Relationship Multiplication

[Jeff Lanfield]

Suppose I have users that can belong to organizations. Organizations
are arranged in a tree. Each organization has only one parent
organization but a user maybe a member of multiple organizations.

The problem that I'm facing that both organizations and individual
users may have relationships with other entities which are
semantically the same. For instance, an individual user can purchase
things and so can an organization. An individual user can have
business partners and so can an organization. So it seems that I would
need to have a duplicate set of link tables that link a user to a
purchase and then a parallel link table linking an organization to a
purchase. If I have N entities with which both users and organizations
may have relationships then I need 2*N link tables. There is nothing
wrong with that per se but just not elegant to have two different
tables for a relationship which is the same in nature, e.g.
purchaser->purchaseditem.

One other approach I was thinking of is to create an intermediate
entity (say it's called "holder") that will be used to hold references
to all the relationships that both an organization and an individual
may have. There will be 2 link tables linking organizations to
"holder" and users to "holder". Holder will in turn reference the
purchases, partners and so on. In this case the number of link tables
will be N+2 as opposed to 2*N but it will have a performance cost of
an extra join.

Is there a better way of modelling this notion of 2 different entities
that can possess similar relationships with N other entities?

[Dawn M. Wolthuis]

"Jeff Lanfield" <jl***********@yahoo.com> wrote in message
news:23**************************@posting.google.c om...

Suppose I have users that can belong to organizations. Organizations
are arranged in a tree. Each organization has only one parent
organization but a user maybe a member of multiple organizations.

The problem that I'm facing that both organizations and individual
users may have relationships with other entities which are
semantically the same. For instance, an individual user can purchase
things and so can an organization. An individual user can have
business partners and so can an organization. So it seems that I would
need to have a duplicate set of link tables that link a user to a
purchase and then a parallel link table linking an organization to a
purchase. If I have N entities with which both users and organizations
may have relationships then I need 2*N link tables. There is nothing
wrong with that per se but just not elegant to have two different
tables for a relationship which is the same in nature, e.g.
purchaser->purchaseditem.

One other approach I was thinking of is to create an intermediate
entity (say it's called "holder") that will be used to hold references
to all the relationships that both an organization and an individual
may have. There will be 2 link tables linking organizations to
"holder" and users to "holder". Holder will in turn reference the
purchases, partners and so on. In this case the number of link tables
will be N+2 as opposed to 2*N but it will have a performance cost of
an extra join.
This is common scenario for a Customer, for example, where the Customer can
then be either an organization or an individual.
Is there a better way of modelling this notion of 2 different entities
that can possess similar relationships with N other entities?

If you are just looking to model it, then using an interface/implementation
approach would work where you have Customer (Holder) as an interface and
Organization and Person both implementing that interface. You could toss in
an abstract class (partial implementation of the Interface) for the
relationship implementations and extend that too. This is more handily
modeled in UML OO types of diagrams than with relations, it seems to me.

But, if you are looking to model it in order to implement in a SQL database
then I think you are headed down the right path with your Holder pattern,
although there could be other approaches that I'm missing too. I'd suggest
putting some work into picking a meaningful name for the Holder relation
(for example, it makes complete sense to everyone that a Customer could be a
person or an org).

Cheers! --dawn

[Alan]

"Jeff Lanfield" <jl***********@yahoo.com> wrote in message
news:23**************************@posting.google.c om...

Suppose I have users that can belong to organizations. Organizations
are arranged in a tree. Each organization has only one parent
organization but a user maybe a member of multiple organizations.

The problem that I'm facing that both organizations and individual
users may have relationships with other entities which are
semantically the same. For instance, an individual user can purchase
things and so can an organization. An individual user can have
business partners and so can an organization. So it seems that I would
need to have a duplicate set of link tables that link a user to a
purchase and then a parallel link table linking an organization to a
purchase. If I have N entities with which both users and organizations
may have relationships then I need 2*N link tables. There is nothing
wrong with that per se but just not elegant to have two different
tables for a relationship which is the same in nature, e.g.
purchaser->purchaseditem.

One other approach I was thinking of is to create an intermediate
entity (say it's called "holder") that will be used to hold references
to all the relationships that both an organization and an individual
may have. There will be 2 link tables linking organizations to
"holder" and users to "holder". Holder will in turn reference the
purchases, partners and so on. In this case the number of link tables
will be N+2 as opposed to 2*N but it will have a performance cost of
an extra join.

Is there a better way of modelling this notion of 2 different entities
that can possess similar relationships with N other entities?

You need to convert the following into an ERD. I've supplied particpation
(may) constraints and cardinality constraints. Once the ERD is done, it is a
snap to convert to a relational schema.

PEOPLE n (may) have m BUSINESS_PARTNERS
ORGANIZATIONS n (may) have BUSINESS_PARTNERS
PEOPLE n (may) belong_to m ORGANIZATIONS
PEOPLE 1 (may) buy n GOODS
ORGANIZATIONS 1 (may) buy n GOODS

So, in mock pseudo ERD form:

PEOPLE == >have m:n ==> BUSINESS_PARTNERS <== n:m have <== ORGANIZATIONS
" ==> people_order ==> 1:m GOODS m:1
<== orgs_order <== "

So, you wind up with

PEOPLE
person_id PK
etc

ORGANIZATIONS
org_id PK
etc

BUSINESS_PARTNERS
person_id PK (FK)
org_id PK (FK)

GOODS
item_id (PK)
description
etc

PEOPLE_ORDER
order_id PK
person_id (FK)
item_id (FK)
order_date
etc

You can now either create an ORGS_ORDER table, or add in org_id to the
PEOPLE_ORDER table, and change the table name to ORDER. This is a decsion
based on many things, primarily the semantics of person_id and org_id. If
they are mutually exclusive, then you can combine the tables simply by
adding org_id. If they are not mutually exclusive, a horrible option is to
add a flag to the ORDERS table, indicating whether the order is from a
person or an organization. Don't do it.

[--CELKO--]

Here is the link on Amazon.com for my new book on "Trees & Hierarchies
in SQL"

http://www.amazon.com/exec/obidos/tg...roduct-details

Separate the tree structure from the nodes. Ilike the nested sets
model for the structure, but you can pick whatever works best for your
situation. Then the nodes can go into another table.

The classic scenario calls for a root class with all the common
attributes and then specialized sub-classes under it. As an example,
let's take the class of Vehicles and find an industry standard
identifier (VIN), and add two mutually exclusive sub-classes, Sport
utility vehicles and sedans ('SUV', 'SED').

CREATE TABLE Vehicles
(vin CHAR(17) NOT NULL PRIMARY KEY,
vehicle_type CHAR(3) NOT NULL
CHECK(vehicle_type IN ('SUV', 'SED')),
UNIQUE (vin, vehicle_type),
..);

Notice the overlapping candidate keys. I then use a compound candidate
key (vin, vehicle_type) and a constraint in each sub-class table to
assure that the vehicle_type is locked and agrees with the Vehicles
table. Add some DRI actions and you are done:

CREATE TABLE SUV
(vin CHAR(17) NOT NULL PRIMARY KEY,
vehicle_type CHAR(3) DEFAULT 'SUV' NOT NULL
CHECK(vehicle_type = 'SUV'),
UNIQUE (vin, vehicle_type),
FOREIGN KEY (vin, vehicle_type)
REFERENCES Vehicles(vin, vehicle_type)
ON UPDATE CASCADE
ON DELETE CASCADE,
..);

CREATE TABLE Sedans
(vin CHAR(17) NOT NULL PRIMARY KEY,
vehicle_type CHAR(3) DEFAULT 'SED' NOT NULL
CHECK(vehicle_type = 'SED'),
UNIQUE (vin, vehicle_type),
FOREIGN KEY (vin, vehicle_type)
REFERENCES Vehicles(vin, vehicle_type)
ON UPDATE CASCADE
ON DELETE CASCADE,
..);

I can continue to build a hierarchy like this. For example, if I had
a Sedans table that broke down into two-door and four-door sedans, I
could a schema like this:

CREATE TABLE Sedans
(vin CHAR(17) NOT NULL PRIMARY KEY,
vehicle_type CHAR(3) DEFAULT 'SED' NOT NULL
CHECK(vehicle_type IN ('2DR', '4DR', 'SED')),
UNIQUE (vin, vehicle_type),
FOREIGN KEY (vin, vehicle_type)
REFERENCES Vehicles(vin, vehicle_type)
ON UPDATE CASCADE
ON DELETE CASCADE,
..);

CREATE TABLE TwoDoor
(vin CHAR(17) NOT NULL PRIMARY KEY,
vehicle_type CHAR(3) DEFAULT '2DR' NOT NULL
CHECK(vehicle_type = '2DR'),
UNIQUE (vin, vehicle_type),
FOREIGN KEY (vin, vehicle_type)
REFERENCES Sedans(vin, vehicle_type)
ON UPDATE CASCADE
ON DELETE CASCADE,
..);

CREATE TABLE FourDoor
(vin CHAR(17) NOT NULL PRIMARY KEY,
vehicle_type CHAR(3) DEFAULT '4DR' NOT NULL
CHECK(vehicle_type = '4DR'),
UNIQUE (vin, vehicle_type),
FOREIGN KEY (vin, vehicle_type)
REFERENCES Sedans (vin, vehicle_type)
ON UPDATE CASCADE
ON DELETE CASCADE,
..);

The idea is to build a chain of identifiers and types in a UNIQUE()
constraint that go up the tree when you use a REFERENCES constraint.
Obviously, you can do variants of this trick to get different class
structures.

If an entity doesn't have to be exclusively one subtype, you play with
the root of the class hierarchy:

CREATE TABLE Vehicles
(vin CHAR(17) NOT NULL,
vehicle_type CHAR(3) NOT NULL
CHECK(vehicle_type IN ('SUV', 'SED')),
PRIMARY KEY (vin, vehicle_type),
..);

Now start hiding all this stuff in VIEWs immediately and add an
INSTEAD OF trigger to those VIEWs.

[John]

--CELKO-- wrote:

Here is the link on Amazon.com for my new book on "Trees & Hierarchies
in SQL"

http://www.amazon.com/exec/obidos/tg...roduct-details

Separate the tree structure from the nodes. Ilike the nested sets
model for the structure, but you can pick whatever works best for your
situation. Then the nodes can go into another table.

<snip a clever way of storing tree structures>

I have a couple of questions:

1. What's to stop you putting extra information into the Sedans or SUV
tables (like descriptions etc.) instead of creating separate tables?

2. Do you think it would be good to have primary keys that span two
tables in order to get rid of the vehicles table? Would there be any
disadvantage to doing it this way instead of the way you outlined?

John

PS: Apologies in advance for (predictably) missing obvious things.

[Joe Celko]

>> 1. What's to stop you putting extra information into the Sedans or
SUV tables (like descriptions etc.) instead of creating separate tables?
<<

If you don't need further subclass breakdowns, then by all means stop at
this level and get all the attributes in the table. In your case,
organizations might have tax status codes which are different from
individuals, so the Customers references the Organzations table and the
Organzations table has a column for tax status which is not in the
Individuals table.

2. Do you think it would be good to have primary keys that span two

tables in order to get rid of the vehicles table? Would there be any
disadvantage to doing it this way instead of the way you outlined? <<

The reason for having a Vehicles table is to establish a class that is
then broken down into disjoint sub-classses; this is the mechanism that
assures a VIN belongs to SUV or Sedans but never both.

What I am proposing is a bit complicated at first sight. You have a
nested set (or whatever) model for the tree structure that might look
like this:

CREATE TABLE Heirarchy
(vin CHAR(17) UNIQUE -- not null? default?
REFERENCES Vehicles (vin)
ON UPDATE CASCADE
ON DELETE ??, -- need a rule
lft INTEGER NOT NULL,
rgt INTEGER NOT NULL,
PRIMARY KEY (lft, rgt),
<<more constraints -- see book>>);

You now have to sit down and really think about the business rules for
the Heirarchy.

The nodes (vehicles) have their class hierarchy on that side of the
RDBMS. You do joins from OrgChart (structure)-> Vehicles (node class)
-> SUV or Sedans (node sub-class) to get your data.

--CELKO--
===========================
Please post DDL, so that people do not have to guess what the keys,
constraints, Declarative Referential Integrity, datatypes, etc. in your
schema are.

*** Sent via Devdex http://www.devdex.com ***
Don't just participate in USENET...get rewarded for it!

[Jeff Lanfield]

> Separate the tree structure from the nodes. Ilike the nested sets

model for the structure, but you can pick whatever works best for your
situation. Then the nodes can go into another table.

The classic scenario calls for a root class with all the common
attributes and then specialized sub-classes under it.

The problem I described was not so much with modelling the tree
(thanks partly to tips from your book which I do own) but with 2
entities having semantically identical relationships with N other
entities. In my case the 2 entities are nodes and leaves -
organizations and users.

For simplicity of the example I'll use the path enumeration model (I
know nested set is better).

create table organizations (int id, varchar name ... etc)
create table users (int id, varchar name ... etc)
create table orgtree (varchar path, int orgId)
create table members (int orgId, int userId) // which organizations a
user belongs to

create table partners (int id, ... etc)
create table purchases (int id, ... etc)
create table customers (int id, ... etc)

The problem is that both users and organizations can have
relationships with partners, purchases, and customers. Thus I have to
have 2*N = 6 link tables to represent that. This example is
simplified, in reality I have about 12 entities with which both
organizations and users can have relationships so I have to have 24
link tables. Nothing wrong with that per se but I just think it is
inelegant.

In OO technology there are standard solutions for this kind of thing
but in SQL the only approach I could think of is to introduce an
"intermediate" entity (say it's called "holder") which will hold the
references to the relationships that both a user and an organization
can have. Users and Organizations can then in turn have a relationship
with this "holder" entity. Thus the number of link tables will be N +
2 or 14 instead of 24.

My question is: Are there any other approaches to this problem? Seems
like a fairly common issue.

Thanks!

Jeff

P.S. CELKO,

I have to deal with tree stuff in SQL a lot and I originally I bought
your book simply because there was nothing else on the topic and I
looked hard both recently and in the past! I did not have high
expectations because I thought it was just a quickie to capitalize on
success of "SQL for smarties" but I was pleasently surprised: as the
title suggests, it is definitely the most comprehensive compilation of
SQL techiques for modelling trees in relational structures all
gathered in one place that I ever saw. Well worth the price,thanks for
putting it together!

jc*******@earthlink.net (--CELKO--) wrote in message news:<18**************************@posting.google. com>... Here is the link on Amazon.com for my new book on "Trees & Hierarchies
in SQL"

http://www.amazon.com/exec/obidos/tg...roduct-details

Separate the tree structure from the nodes. Ilike the nested sets
model for the structure, but you can pick whatever works best for your
situation. Then the nodes can go into another table.

The classic scenario calls for a root class with all the common
attributes and then specialized sub-classes under it. As an example,
let's take the class of Vehicles and find an industry standard
identifier (VIN), and add two mutually exclusive sub-classes, Sport
utility vehicles and sedans ('SUV', 'SED').

CREATE TABLE Vehicles
(vin CHAR(17) NOT NULL PRIMARY KEY,
vehicle_type CHAR(3) NOT NULL
CHECK(vehicle_type IN ('SUV', 'SED')),
UNIQUE (vin, vehicle_type),
..);

Notice the overlapping candidate keys. I then use a compound candidate
key (vin, vehicle_type) and a constraint in each sub-class table to
assure that the vehicle_type is locked and agrees with the Vehicles
table. Add some DRI actions and you are done:

CREATE TABLE SUV
(vin CHAR(17) NOT NULL PRIMARY KEY,
vehicle_type CHAR(3) DEFAULT 'SUV' NOT NULL
CHECK(vehicle_type = 'SUV'),
UNIQUE (vin, vehicle_type),
FOREIGN KEY (vin, vehicle_type)
REFERENCES Vehicles(vin, vehicle_type)
ON UPDATE CASCADE
ON DELETE CASCADE,
..);

CREATE TABLE Sedans
(vin CHAR(17) NOT NULL PRIMARY KEY,
vehicle_type CHAR(3) DEFAULT 'SED' NOT NULL
CHECK(vehicle_type = 'SED'),
UNIQUE (vin, vehicle_type),
FOREIGN KEY (vin, vehicle_type)
REFERENCES Vehicles(vin, vehicle_type)
ON UPDATE CASCADE
ON DELETE CASCADE,
..);

I can continue to build a hierarchy like this. For example, if I had
a Sedans table that broke down into two-door and four-door sedans, I
could a schema like this:

CREATE TABLE Sedans
(vin CHAR(17) NOT NULL PRIMARY KEY,
vehicle_type CHAR(3) DEFAULT 'SED' NOT NULL
CHECK(vehicle_type IN ('2DR', '4DR', 'SED')),
UNIQUE (vin, vehicle_type),
FOREIGN KEY (vin, vehicle_type)
REFERENCES Vehicles(vin, vehicle_type)
ON UPDATE CASCADE
ON DELETE CASCADE,
..);

CREATE TABLE TwoDoor
(vin CHAR(17) NOT NULL PRIMARY KEY,
vehicle_type CHAR(3) DEFAULT '2DR' NOT NULL
CHECK(vehicle_type = '2DR'),
UNIQUE (vin, vehicle_type),
FOREIGN KEY (vin, vehicle_type)
REFERENCES Sedans(vin, vehicle_type)
ON UPDATE CASCADE
ON DELETE CASCADE,
..);

CREATE TABLE FourDoor
(vin CHAR(17) NOT NULL PRIMARY KEY,
vehicle_type CHAR(3) DEFAULT '4DR' NOT NULL
CHECK(vehicle_type = '4DR'),
UNIQUE (vin, vehicle_type),
FOREIGN KEY (vin, vehicle_type)
REFERENCES Sedans (vin, vehicle_type)
ON UPDATE CASCADE
ON DELETE CASCADE,
..);

The idea is to build a chain of identifiers and types in a UNIQUE()
constraint that go up the tree when you use a REFERENCES constraint.
Obviously, you can do variants of this trick to get different class
structures.

If an entity doesn't have to be exclusively one subtype, you play with
the root of the class hierarchy:

CREATE TABLE Vehicles
(vin CHAR(17) NOT NULL,
vehicle_type CHAR(3) NOT NULL
CHECK(vehicle_type IN ('SUV', 'SED')),
PRIMARY KEY (vin, vehicle_type),
..);

Now start hiding all this stuff in VIEWs immediately and add an
INSTEAD OF trigger to those VIEWs.

[Jeff Lanfield]

> Separate the tree structure from the nodes. Ilike the nested sets

model for the structure, but you can pick whatever works best for your
situation. Then the nodes can go into another table.
CELKO,

Just wanted to double check: by "separating the nodes from the tree"
do you mean doing this (using path enumeration model):

nodes (id int, name varchar, ... etc)
nodetree (path varchar, int nodeId)

AS OPPOSED TO THIS:
nodes (id int, name varchar, path varchar, ... etc)

Thanks again!
Jeff
jc*******@earthlink.net (--CELKO--) wrote in message news:<18**************************@posting.google. com>... Here is the link on Amazon.com for my new book on "Trees & Hierarchies
in SQL"

http://www.amazon.com/exec/obidos/tg...roduct-details

Separate the tree structure from the nodes. Ilike the nested sets
model for the structure, but you can pick whatever works best for your
situation. Then the nodes can go into another table.

The classic scenario calls for a root class with all the common
attributes and then specialized sub-classes under it. As an example,
let's take the class of Vehicles and find an industry standard
identifier (VIN), and add two mutually exclusive sub-classes, Sport
utility vehicles and sedans ('SUV', 'SED').

CREATE TABLE Vehicles
(vin CHAR(17) NOT NULL PRIMARY KEY,
vehicle_type CHAR(3) NOT NULL
CHECK(vehicle_type IN ('SUV', 'SED')),
UNIQUE (vin, vehicle_type),
..);

Notice the overlapping candidate keys. I then use a compound candidate
key (vin, vehicle_type) and a constraint in each sub-class table to
assure that the vehicle_type is locked and agrees with the Vehicles
table. Add some DRI actions and you are done:

CREATE TABLE SUV
(vin CHAR(17) NOT NULL PRIMARY KEY,
vehicle_type CHAR(3) DEFAULT 'SUV' NOT NULL
CHECK(vehicle_type = 'SUV'),
UNIQUE (vin, vehicle_type),
FOREIGN KEY (vin, vehicle_type)
REFERENCES Vehicles(vin, vehicle_type)
ON UPDATE CASCADE
ON DELETE CASCADE,
..);

CREATE TABLE Sedans
(vin CHAR(17) NOT NULL PRIMARY KEY,
vehicle_type CHAR(3) DEFAULT 'SED' NOT NULL
CHECK(vehicle_type = 'SED'),
UNIQUE (vin, vehicle_type),
FOREIGN KEY (vin, vehicle_type)
REFERENCES Vehicles(vin, vehicle_type)
ON UPDATE CASCADE
ON DELETE CASCADE,
..);

I can continue to build a hierarchy like this. For example, if I had
a Sedans table that broke down into two-door and four-door sedans, I
could a schema like this:

CREATE TABLE Sedans
(vin CHAR(17) NOT NULL PRIMARY KEY,
vehicle_type CHAR(3) DEFAULT 'SED' NOT NULL
CHECK(vehicle_type IN ('2DR', '4DR', 'SED')),
UNIQUE (vin, vehicle_type),
FOREIGN KEY (vin, vehicle_type)
REFERENCES Vehicles(vin, vehicle_type)
ON UPDATE CASCADE
ON DELETE CASCADE,
..);

CREATE TABLE TwoDoor
(vin CHAR(17) NOT NULL PRIMARY KEY,
vehicle_type CHAR(3) DEFAULT '2DR' NOT NULL
CHECK(vehicle_type = '2DR'),
UNIQUE (vin, vehicle_type),
FOREIGN KEY (vin, vehicle_type)
REFERENCES Sedans(vin, vehicle_type)
ON UPDATE CASCADE
ON DELETE CASCADE,
..);

CREATE TABLE FourDoor
(vin CHAR(17) NOT NULL PRIMARY KEY,
vehicle_type CHAR(3) DEFAULT '4DR' NOT NULL
CHECK(vehicle_type = '4DR'),
UNIQUE (vin, vehicle_type),
FOREIGN KEY (vin, vehicle_type)
REFERENCES Sedans (vin, vehicle_type)
ON UPDATE CASCADE
ON DELETE CASCADE,
..);

The idea is to build a chain of identifiers and types in a UNIQUE()
constraint that go up the tree when you use a REFERENCES constraint.
Obviously, you can do variants of this trick to get different class
structures.

If an entity doesn't have to be exclusively one subtype, you play with
the root of the class hierarchy:

CREATE TABLE Vehicles
(vin CHAR(17) NOT NULL,
vehicle_type CHAR(3) NOT NULL
CHECK(vehicle_type IN ('SUV', 'SED')),
PRIMARY KEY (vin, vehicle_type),
..);

Now start hiding all this stuff in VIEWs immediately and add an
INSTEAD OF trigger to those VIEWs.

[Joe Celko]

>> in reality I have about 12 entities with which both
organizations and users can have relationships so I have to have 24 link
tables. Nothing wrong with that per se but I just think it is inelegant.
<<

Well, sometimes the model is complex. If those relationships are all
different, then you need to have table for each relationship.

I did not have high expectations because I thought it was just a

quickie to capitalize on success of "SQL for smarties" but I was
pleasantly surprised: as the title suggests, it is definitely the most
comprehensive compilation of SQL techiques for modelling trees in
relational structures all gathered in one place that I ever saw. <<

Thank you! Where were you when I needed jacket copy?

--CELKO--
===========================
Please post DDL, so that people do not have to guess what the keys,
constraints, Declarative Referential Integrity, datatypes, etc. in your
schema are.

*** Sent via Devdex http://www.devdex.com ***
Don't just participate in USENET...get rewarded for it!

[--CELKO--]

>> Just wanted to double check: by "separating the nodes from the
tree" do you mean doing this (using path enumeration model): <<

Where is the DDL? What you posted was useless on several levels.
NEVER use something as vague as "id" for a column name. NEVER, NEVER,
NEVER name a data element for its current location -- node_id always a
node_id wherever it appears. This is fundamental data modeling --
name thing s for what they are in the logical model, not for how or
where they are PHYSICALLY stored or how they are used in one place.

I think that you might have meant this:

CREATE TABLE Nodes
(node_id INTEGER NOT NULL PRIMARY KEY,
node_name CHAR(20) NOT NULL,
... );

CREATE TABLE Tree
(node_id INTEGER NOT NULL
REFERENCES Nodes (node_id)
ON UPDATE CASCADE
ON DELETE CASCADE,
path VARCHAR(100) NOT NULL
CHECK (<< valid path predicate >>),
.. );

Now you need to decide how to handle DRI actions; I made a guess. And
how you want to build the path string - letters, digits, fixed or
variable substring components, etc.

[Daniel Morgan]

--CELKO-- wrote:

Here is the link on Amazon.com for my new book on "Trees & Hierarchies
in SQL"

http://www.amazon.com/exec/obidos/tg...roduct-details

Separate the tree structure from the nodes. Ilike the nested sets
model for the structure, but you can pick whatever works best for your
situation. Then the nodes can go into another table.

The classic scenario calls for a root class with all the common
attributes and then specialized sub-classes under it. As an example,
let's take the class of Vehicles and find an industry standard
identifier (VIN), and add two mutually exclusive sub-classes, Sport
utility vehicles and sedans ('SUV', 'SED').

CREATE TABLE Vehicles
(vin CHAR(17) NOT NULL PRIMARY KEY,
vehicle_type CHAR(3) NOT NULL
CHECK(vehicle_type IN ('SUV', 'SED')),
UNIQUE (vin, vehicle_type),
..);

Notice the overlapping candidate keys. I then use a compound candidate
key (vin, vehicle_type) and a constraint in each sub-class table to
assure that the vehicle_type is locked and agrees with the Vehicles
table. Add some DRI actions and you are done:

CREATE TABLE SUV
(vin CHAR(17) NOT NULL PRIMARY KEY,
vehicle_type CHAR(3) DEFAULT 'SUV' NOT NULL
CHECK(vehicle_type = 'SUV'),
UNIQUE (vin, vehicle_type),
FOREIGN KEY (vin, vehicle_type)
REFERENCES Vehicles(vin, vehicle_type)
ON UPDATE CASCADE
ON DELETE CASCADE,
..);

CREATE TABLE Sedans
(vin CHAR(17) NOT NULL PRIMARY KEY,
vehicle_type CHAR(3) DEFAULT 'SED' NOT NULL
CHECK(vehicle_type = 'SED'),
UNIQUE (vin, vehicle_type),
FOREIGN KEY (vin, vehicle_type)
REFERENCES Vehicles(vin, vehicle_type)
ON UPDATE CASCADE
ON DELETE CASCADE,
..);

I can continue to build a hierarchy like this. For example, if I had
a Sedans table that broke down into two-door and four-door sedans, I
could a schema like this:

CREATE TABLE Sedans
(vin CHAR(17) NOT NULL PRIMARY KEY,
vehicle_type CHAR(3) DEFAULT 'SED' NOT NULL
CHECK(vehicle_type IN ('2DR', '4DR', 'SED')),
UNIQUE (vin, vehicle_type),
FOREIGN KEY (vin, vehicle_type)
REFERENCES Vehicles(vin, vehicle_type)
ON UPDATE CASCADE
ON DELETE CASCADE,
..);

CREATE TABLE TwoDoor
(vin CHAR(17) NOT NULL PRIMARY KEY,
vehicle_type CHAR(3) DEFAULT '2DR' NOT NULL
CHECK(vehicle_type = '2DR'),
UNIQUE (vin, vehicle_type),
FOREIGN KEY (vin, vehicle_type)
REFERENCES Sedans(vin, vehicle_type)
ON UPDATE CASCADE
ON DELETE CASCADE,
..);

CREATE TABLE FourDoor
(vin CHAR(17) NOT NULL PRIMARY KEY,
vehicle_type CHAR(3) DEFAULT '4DR' NOT NULL
CHECK(vehicle_type = '4DR'),
UNIQUE (vin, vehicle_type),
FOREIGN KEY (vin, vehicle_type)
REFERENCES Sedans (vin, vehicle_type)
ON UPDATE CASCADE
ON DELETE CASCADE,
..);

The idea is to build a chain of identifiers and types in a UNIQUE()
constraint that go up the tree when you use a REFERENCES constraint.
Obviously, you can do variants of this trick to get different class
structures.

If an entity doesn't have to be exclusively one subtype, you play with
the root of the class hierarchy:

CREATE TABLE Vehicles
(vin CHAR(17) NOT NULL,
vehicle_type CHAR(3) NOT NULL
CHECK(vehicle_type IN ('SUV', 'SED')),
PRIMARY KEY (vin, vehicle_type),
..);

Now start hiding all this stuff in VIEWs immediately and add an
INSTEAD OF trigger to those VIEWs.

Joe ... sorry ... but integrity demands that I write the following:

We have a usenet group named comp.databases.oracle.marketplace
specifically designated for promotions. In the future it would be
appreciated if you posted book, or any other, promotions there.

Thanks.

For everyone else ... I recommend Joe's books to my students and
highly recommend them.

--
Daniel Morgan
http://www.outreach.washington.edu/e...ad/oad_crs.asp
http://www.outreach.washington.edu/e...oa/aoa_crs.asp
da******@x.washington.edu
(replace 'x' with a 'u' to reply)

[Marshall Spight]

"Daniel Morgan" <da******@x.washington.edu> wrote in message news:1088831855.922891@yasure...

Joe ... sorry ... but integrity demands that I write the following:

We have a usenet group named comp.databases.oracle.marketplace
specifically designated for promotions. In the future it would be
appreciated if you posted book, or any other, promotions there.

Uh, so a guy posts two lines with a link where you can buy his
book, and 100 lines of helping someone out with his question,
and you bust his balls for "promotions?" You also assume
that because the original poster included an oracle newsgroup
in one of the four newsgroups he posted to, that everyone responding
on any of those newsgroups should somehow know your local
culture and customs and adhere to them?

You have a funny idea of integrity.
Marshall

[Daniel Morgan]

Marshall Spight wrote:

"Daniel Morgan" <da******@x.washington.edu> wrote in message news:1088831855.922891@yasure...

Joe ... sorry ... but integrity demands that I write the following:

We have a usenet group named comp.databases.oracle.marketplace
specifically designated for promotions. In the future it would be
appreciated if you posted book, or any other, promotions there.

Uh, so a guy posts two lines with a link where you can buy his
book, and 100 lines of helping someone out with his question,
and you bust his balls for "promotions?" You also assume
that because the original poster included an oracle newsgroup
in one of the four newsgroups he posted to, that everyone responding
on any of those newsgroups should somehow know your local
culture and customs and adhere to them?

You have a funny idea of integrity.
Marshall

I don't show favoritism when it come to calling spam spam. Even when it
is someone as esteemed as Joe Celko.

But exactly what is it you have contributed to c.d.o.server?

--
Daniel Morgan
http://www.outreach.washington.edu/e...ad/oad_crs.asp
http://www.outreach.washington.edu/e...oa/aoa_crs.asp
da******@x.washington.edu
(replace 'x' with a 'u' to reply)

[Gene Wirchenko]

Daniel Morgan <da******@x.washington.edu> wrote:

Marshall Spight wrote:

"Daniel Morgan" <da******@x.washington.edu> wrote in message news:1088831855.922891@yasure...

Joe ... sorry ... but integrity demands that I write the following:

We have a usenet group named comp.databases.oracle.marketplace
specifically designated for promotions. In the future it would be
appreciated if you posted book, or any other, promotions there.

Uh, so a guy posts two lines with a link where you can buy his
book, and 100 lines of helping someone out with his question,
and you bust his balls for "promotions?" You also assume
that because the original poster included an oracle newsgroup
in one of the four newsgroups he posted to, that everyone responding
on any of those newsgroups should somehow know your local
culture and customs and adhere to them?

You have a funny idea of integrity.

I don't show favoritism when it come to calling spam spam. Even when it
is someone as esteemed as Joe Celko.
What spam? OP asked for help, and Mr. Celko gave it.
But exactly what is it you have contributed to c.d.o.server?

Possibly a reminder of priorities? You may have missed this point
though.

Sincerely,

Gene Wirchenko

Computerese Irregular Verb Conjugation:
I have preferences.
You have biases.
He/She has prejudices.

[Marshall Spight]

"Daniel Morgan" <da******@x.washington.edu> wrote in message news:1088887202.344807@yasure...

I don't show favoritism when it come to calling spam spam. Even when it
is someone as esteemed as Joe Celko.
Spam is by definition unsolicited. An on-topic, helpful response to a
usenet post requesting help is by definition solicited. Your "integrity"
is just ball-busting on a guy who bends over backwards to help all
comers for free, and also happens to sell some books for which
he's probably lucky to clear $25,000 annually. (Trees and
Hierarchies in SQL doesn't have the same cachet as the latest
work by David Sedaris. In fact, it's Amazon's 16,754's most
popular item, right behind the $350 "Suunto Yachtsman Wristop
Computer Watch w/ Barometer and Compass", and you can
bet a $350 wrist barometer isn't exactly flying off the shelves.)

You're not engaging in spam-vigilance; you're just engaging
in gratuitous anticommercialism.

But exactly what is it you have contributed to c.d.o.server?

I could just as well ask what you've contributed to c.d.theory,
but that wasn't the point. The point was that this thread isn't
specific to your newsgroup; it spans multiple newsgroups.
Expecting someone on this thread to conform to the customs
of one of those newsgroups is as unrealistic as those folks in
Snakewater, Iowa, who get upset when they see something
that violates a local ordinance posted on the internet from
New York City.

But the absolute frosting on the cake, the hypocritical parsley
on the potatoes, is the fact that both of your posts on this
topic so far have included links to courses you teach at
Washington. You're giving those courses away for free,
I presume? Oh, no, wait, I see; Oracle Application Development
is $1875. At least Joe's links to his book on Amazon were
pertinent to the OP's question.
Marshall

[omlet]

Daniel Morgan <da******@x.washington.edu> wrote in message news:<1088831855.922891@yasure>...

--CELKO-- wrote:

Here is the link on Amazon.com for my new book on "Trees & Hierarchies
in SQL"

http://www.amazon.com/exec/obidos/tg...roduct-details

Separate the tree structure from the nodes. Ilike the nested sets
model for the structure, but you can pick whatever works best for your
situation. Then the nodes can go into another table.

The classic scenario calls for a root class with all the common
attributes and then specialized sub-classes under it. As an example,
let's take the class of Vehicles and find an industry standard
identifier (VIN), and add two mutually exclusive sub-classes, Sport
utility vehicles and sedans ('SUV', 'SED').

CREATE TABLE Vehicles
(vin CHAR(17) NOT NULL PRIMARY KEY,
vehicle_type CHAR(3) NOT NULL
CHECK(vehicle_type IN ('SUV', 'SED')),
UNIQUE (vin, vehicle_type),
..);

Notice the overlapping candidate keys. I then use a compound candidate
key (vin, vehicle_type) and a constraint in each sub-class table to
assure that the vehicle_type is locked and agrees with the Vehicles
table. Add some DRI actions and you are done:

CREATE TABLE SUV
(vin CHAR(17) NOT NULL PRIMARY KEY,
vehicle_type CHAR(3) DEFAULT 'SUV' NOT NULL
CHECK(vehicle_type = 'SUV'),
UNIQUE (vin, vehicle_type),
FOREIGN KEY (vin, vehicle_type)
REFERENCES Vehicles(vin, vehicle_type)
ON UPDATE CASCADE
ON DELETE CASCADE,
..);

CREATE TABLE Sedans
(vin CHAR(17) NOT NULL PRIMARY KEY,
vehicle_type CHAR(3) DEFAULT 'SED' NOT NULL
CHECK(vehicle_type = 'SED'),
UNIQUE (vin, vehicle_type),
FOREIGN KEY (vin, vehicle_type)
REFERENCES Vehicles(vin, vehicle_type)
ON UPDATE CASCADE
ON DELETE CASCADE,
..);

I can continue to build a hierarchy like this. For example, if I had
a Sedans table that broke down into two-door and four-door sedans, I
could a schema like this:

CREATE TABLE Sedans
(vin CHAR(17) NOT NULL PRIMARY KEY,
vehicle_type CHAR(3) DEFAULT 'SED' NOT NULL
CHECK(vehicle_type IN ('2DR', '4DR', 'SED')),
UNIQUE (vin, vehicle_type),
FOREIGN KEY (vin, vehicle_type)
REFERENCES Vehicles(vin, vehicle_type)
ON UPDATE CASCADE
ON DELETE CASCADE,
..);

CREATE TABLE TwoDoor
(vin CHAR(17) NOT NULL PRIMARY KEY,
vehicle_type CHAR(3) DEFAULT '2DR' NOT NULL
CHECK(vehicle_type = '2DR'),
UNIQUE (vin, vehicle_type),
FOREIGN KEY (vin, vehicle_type)
REFERENCES Sedans(vin, vehicle_type)
ON UPDATE CASCADE
ON DELETE CASCADE,
..);

CREATE TABLE FourDoor
(vin CHAR(17) NOT NULL PRIMARY KEY,
vehicle_type CHAR(3) DEFAULT '4DR' NOT NULL
CHECK(vehicle_type = '4DR'),
UNIQUE (vin, vehicle_type),
FOREIGN KEY (vin, vehicle_type)
REFERENCES Sedans (vin, vehicle_type)
ON UPDATE CASCADE
ON DELETE CASCADE,
..);

The idea is to build a chain of identifiers and types in a UNIQUE()
constraint that go up the tree when you use a REFERENCES constraint.
Obviously, you can do variants of this trick to get different class
structures.

If an entity doesn't have to be exclusively one subtype, you play with
the root of the class hierarchy:

CREATE TABLE Vehicles
(vin CHAR(17) NOT NULL,
vehicle_type CHAR(3) NOT NULL
CHECK(vehicle_type IN ('SUV', 'SED')),
PRIMARY KEY (vin, vehicle_type),
..);

Now start hiding all this stuff in VIEWs immediately and add an
INSTEAD OF trigger to those VIEWs.

Joe ... sorry ... but integrity demands that I write the following:

We have a usenet group named comp.databases.oracle.marketplace
specifically designated for promotions. In the future it would be
appreciated if you posted book, or any other, promotions there.

Thanks.

For everyone else ... I recommend Joe's books to my students and
highly recommend them.

Daniel Morgan: Cannot you just shut the ffff up? Take your
contributions and shove it! This an international newsgroup and it NOT
yours to guard. The charter of this newsgroup is clear and does not
allow you to add shameless lines of advertisements while denying
others!

So shut up! you can create your own newsgroup where you can control
things --- you crazy control freak. Watch the movie "Tow Truck" to see
what you look like.

Obviously you are downgrading this group and many others to your level
of integrity which you completely lack.

Remember, you big jumble of Q that Oracle itself is a product! and you
have no rights what so ever to determine what is good; and what is
like you of what you don't like. You get it?! Read Simon's peom! Or
download OMLET and run it and you would get random lines of insults
and images geared specifically to YOU! Till 2010.

Now what?!
Right on your tail.

[Noons]

om***@omlet.org apparently said,on my timestamp of 5/07/2004 8:27 AM:

Daniel Morgan: Cannot you just shut the ffff up? Take your
contributions and shove it! This an international newsgroup and it NOT
yours to guard. The charter of this newsgroup is clear and does not
allow you to add shameless lines of advertisements while denying
others!

So shut up! you can create your own newsgroup where you can control
things --- you crazy control freak. Watch the movie "Tow Truck" to see
what you look like.

Obviously you are downgrading this group and many others to your level
of integrity which you completely lack.

Remember, you big jumble of Q that Oracle itself is a product! and you
have no rights what so ever to determine what is good; and what is
like you of what you don't like. You get it?! Read Simon's peom! Or
download OMLET and run it and you would get random lines of insults
and images geared specifically to YOU! Till 2010.

Now what?!
Right on your tail.

forwarded to abuse.

--
Cheers
Nuno Souto
wi*******@yahoo.com.au.nospam

[Daniel Morgan]

Noons wrote:

forwarded to abuse.

I have too as have a number of my students.

Here's what we have found:

Pinging omlet.org returns 66.218.79.158 which in return, when pinged,
returns p2w2.geo.scd.yahoo.com.

Complaints have been filed with:
ab***@yahoo.com
gr**********@google.com

I encourage others to do so too and and in case anyone wonders who this
spam-troll really is ...

Domain Name:OMLET.ORG
Created On:15-Jun-2004 10:54:38 UTC ... (been in business only 2 weeks)
Sponsoring Registrar:R52-LROR
Registrant Name:Amjad Daoud
Registrant Street1:Queen Noor Br 07
Registrant City:Amman
Registrant State/Province:Amman
Registrant Postal Code:9626
Registrant Country:JO ... (this is the country of Jordan)
Registrant Phone:[phone]
Registrant Email:te******************@yahoo.com

Yes our troll lives in the country of Jordan. You can purchase a
product from someone to use with your valuable Oracle database
where there is not only no company behind it there is just one
guy in a third-world country.

Please address complaints about his spamming and abuse to:
Tech Name:YahooDomains TechContact
Tech Organization:Yahoo! Inc
Tech Street1:701 First Ave.
Tech City:Sunnyvale
Tech State/Province:CA
Tech Postal Code:94089
Tech Country:US
Tech Phone:[phone]
Tech Email:do*********@YAHOO-INC.COM
Name Server:YNS1.YAHOO.COM
Name Server:YNS2.YAHOO.COM

Thank you.
--
Daniel Morgan
http://www.outreach.washington.edu/e...ad/oad_crs.asp
http://www.outreach.washington.edu/e...oa/aoa_crs.asp
da******@x.washington.edu
(replace 'x' with a 'u' to reply)

[x]

**** Post for FREE via your newsreader at post.usenet.com ****
"Daniel Morgan" <da******@x.washington.edu> wrote in message
news:1089057907.319050@yasure...

Here's what we have found: I encourage others to do so too and and in case anyone wonders who this
spam-troll really is ... Domain Name:OMLET.ORG
Created On:15-Jun-2004 10:54:38 UTC ... (been in business only 2 weeks)
Sponsoring Registrar:R52-LROR
Registrant Name:Amjad Daoud
Registrant Street1:Queen Noor Br 07
Registrant City:Amman
Registrant State/Province:Amman
Registrant Postal Code:9626
Registrant Country:JO ... (this is the country of Jordan)
Registrant Phone:[phone]
Registrant Email:te******************@yahoo.com Yes our troll lives in the country of Jordan. You can purchase a
product from someone to use with your valuable Oracle database
where there is not only no company behind it there is just one
guy in a third-world country.

You need to be more careful about this kind of language.
One of Amjad Daoud signatures I found on the web is:

Amjad Daoud
The OMLET Team Lead
Tera Knowledge Systems, Inc.
Arlington, Texas <--------------- This is not Jordan

Besides: What do you have against "guys from third-world countries" ?


-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
*** Usenet.com - The #1 Usenet Newsgroup Service on The Planet! ***
http://www.usenet.com
Unlimited Download - 19 Seperate Servers - 90,000 groups - Uncensored
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=

[Jeff Lanfield]

As the original poster I would agree here. If it is solicited and
relevant to the question asked it's not spam. I actually did end up
buying Mr. Celko's book and can attest that it is quite useful for
solving the class of problems I was interested in.

While in general people should not post ads for books they wrote on
tech newsgroups an exception should definitely be made for cases like
this one where the book contains pretty much the exact answer.

- Jeff
"Marshall Spight" <ms*****@dnai.com> wrote in message news:<8KNFc.20603$%_6.12831@attbi_s01>...

"Daniel Morgan" <da******@x.washington.edu> wrote in message news:1088887202.344807@yasure...

I don't show favoritism when it come to calling spam spam. Even when it
is someone as esteemed as Joe Celko.

Spam is by definition unsolicited. An on-topic, helpful response to a
usenet post requesting help is by definition solicited. Your "integrity"
is just ball-busting on a guy who bends over backwards to help all
comers for free, and also happens to sell some books for which
he's probably lucky to clear $25,000 annually. (Trees and
Hierarchies in SQL doesn't have the same cachet as the latest
work by David Sedaris. In fact, it's Amazon's 16,754's most
popular item, right behind the $350 "Suunto Yachtsman Wristop
Computer Watch w/ Barometer and Compass", and you can
bet a $350 wrist barometer isn't exactly flying off the shelves.)

You're not engaging in spam-vigilance; you're just engaging
in gratuitous anticommercialism.

But exactly what is it you have contributed to c.d.o.server?

I could just as well ask what you've contributed to c.d.theory,
but that wasn't the point. The point was that this thread isn't
specific to your newsgroup; it spans multiple newsgroups.
Expecting someone on this thread to conform to the customs
of one of those newsgroups is as unrealistic as those folks in
Snakewater, Iowa, who get upset when they see something
that violates a local ordinance posted on the internet from
New York City.

But the absolute frosting on the cake, the hypocritical parsley
on the potatoes, is the fact that both of your posts on this
topic so far have included links to courses you teach at
Washington. You're giving those courses away for free,
I presume? Oh, no, wait, I see; Oracle Application Development
is $1875. At least Joe's links to his book on Amazon were
pertinent to the OP's question.
Marshall

[Mikito Harakiri]

"Marshall Spight" <ms*****@dnai.com> wrote in message
news:8KNFc.20603$%_6.12831@attbi_s01...

...Trees and
Hierarchies in SQL doesn't have the same cachet as the latest
work by David Sedaris. In fact, it's Amazon's 16,754's most
popular item, right behind the $350 "Suunto Yachtsman Wristop
Computer Watch w/ Barometer and Compass...

How did you query this? Does amazon provides SQL interface nowadays?

select book.* from books
where rank between 15000 and 16000

???

[Ed prochak]

"x" <x-*****@yahoo.com> wrote in message news:<40******@post.usenet.com>...

**** Post for FREE via your newsreader at post.usenet.com ****
"Daniel Morgan" <da******@x.washington.edu> wrote in message
news:1089057907.319050@yasure...

Here's what we have found:

I encourage others to do so too and and in case anyone wonders who this
spam-troll really is ...

Domain Name:OMLET.ORG
Created On:15-Jun-2004 10:54:38 UTC ... (been in business only 2 weeks)
Sponsoring Registrar:R52-LROR
Registrant Name:Amjad Daoud
Registrant Street1:Queen Noor Br 07
Registrant City:Amman
Registrant State/Province:Amman
Registrant Postal Code:9626
Registrant Country:JO ... (this is the country of Jordan)
Registrant Phone:[phone]
Registrant Email:te******************@yahoo.com

Yes our troll lives in the country of Jordan. You can purchase a
product from someone to use with your valuable Oracle database
where there is not only no company behind it there is just one
guy in a third-world country.

You need to be more careful about this kind of language.
One of Amjad Daoud signatures I found on the web is:

Amjad Daoud
The OMLET Team Lead
Tera Knowledge Systems, Inc.
Arlington, Texas <--------------- This is not Jordan

Besides: What do you have against "guys from third-world countries" ?

He doesn't. Daniel, I and others have just been dealing with this
troll in the ORACLE groups for a while now. And you should note that
you presented the company address, which may be only a sales office
and not the location of the "OMLET Team". It certainly is different
from the residence of this guy.

Calling Amjad a troll is certainly MUCH more polite than the crap he
has spewed in the ORACLE groups.

[Joel Garry]

"x" <x-*****@yahoo.com> wrote in message news:<40******@post.usenet.com>...

**** Post for FREE via your newsreader at post.usenet.com ****
"Daniel Morgan" <da******@x.washington.edu> wrote in message
news:1089057907.319050@yasure...

Here's what we have found:

I encourage others to do so too and and in case anyone wonders who this
spam-troll really is ...

Domain Name:OMLET.ORG
Created On:15-Jun-2004 10:54:38 UTC ... (been in business only 2 weeks)
Sponsoring Registrar:R52-LROR
Registrant Name:Amjad Daoud
Registrant Street1:Queen Noor Br 07
Registrant City:Amman
Registrant State/Province:Amman
Registrant Postal Code:9626
Registrant Country:JO ... (this is the country of Jordan)
Registrant Phone:[phone]
Registrant Email:te******************@yahoo.com

Yes our troll lives in the country of Jordan. You can purchase a
product from someone to use with your valuable Oracle database
where there is not only no company behind it there is just one
guy in a third-world country.

You need to be more careful about this kind of language.
One of Amjad Daoud signatures I found on the web is:

Amjad Daoud
The OMLET Team Lead
Tera Knowledge Systems, Inc.
Arlington, Texas <--------------- This is not Jordan

But check out eggy-weggies email, it's Jordanian. Not that that means
anything, I certainly am not in Tonga or oz as some of my web presence
may imply. But he says Sunnyvale, and Texas, and the registration is
on Queen Noor's... something. All-in-all, not a situation a rational
person would want to take advice or a product from. And wassup with
the 619 area code on the Sunnyvale provider, a cell?

Besides: What do you have against "guys from third-world countries" ?

There are matters of security, economics, legal liability and trust
involved.

But besides that, rotten egg seems to have gone well into
net-psychoses, cyberstalking anyone who he perceives as attacking him,
which is pretty much anyone acknowledging his existence. Not a pretty
sight. DM has once again made a fast, correct call about a spammer.

As far as celko, I thought DM was suitably polite in explaining the
plug issue. It was a minor gaffe on celko's part, perfectly
understandable, DM didn't "break his balls." I would say DM should
have ignored it giving the crossposting, but he clearly is making an
effort to be fair and non-partisan when netcopping. My bias is
towards letting people plug their work amid a useful answer, but DM
has almost convinced me his way is better.

jg
--
@home.com is bogus.
Internet meltdown, news at 11:
http://catless.ncl.ac.uk/Risks/23.44.html#subj1

[Gene Wirchenko]

jl***********@yahoo.com (Jeff Lanfield) wrote:

As the original poster I would agree here. If it is solicited and
relevant to the question asked it's not spam. I actually did end up
buying Mr. Celko's book and can attest that it is quite useful for
solving the class of problems I was interested in.

While in general people should not post ads for books they wrote on
tech newsgroups an exception should definitely be made for cases like
this one where the book contains pretty much the exact answer.

Exactly.

I am glad you got the help you needed. That is a large part of
what technical newsgroups are for.

[snip]

Sincerely,

Gene Wirchenko

Computerese Irregular Verb Conjugation:
I have preferences.
You have biases.
He/She has prejudices.

[Marshall Spight]

"Mikito Harakiri" <mi*********@iahu.com> wrote in message news:wI**************@news.oracle.com...

"Marshall Spight" <ms*****@dnai.com> wrote in message
news:8KNFc.20603$%_6.12831@attbi_s01...

...Trees and
Hierarchies in SQL doesn't have the same cachet as the latest
work by David Sedaris. In fact, it's Amazon's 16,754's most
popular item, right behind the $350 "Suunto Yachtsman Wristop
Computer Watch w/ Barometer and Compass...

How did you query this? Does amazon provides SQL interface nowadays?

Methodology:

Click on the link provided by Mr. Celko.
Scroll down to the "product details" section.
Read the sales rank there (note that it's changed since then.)
--> "Amazon.com sales rank: 16,754"
subtract one :-)
--> 16,753
Go to www.google.com
search exactly as follows:
site:amazon.com "sales rank: 16753"
Click on the link

Often when a site doesn't provide some bit of
info directly, there's a way to get to it with
judicious use of google features.

For example, if you want to know if a given
32 bit number is prime, often times the fastest
way to decide is to google for it. If it's prime,
it'll be on some indexed page of prime numbers
somewhere.
Marshall

[x]

**** Post for FREE via your newsreader at post.usenet.com ****
"Ed prochak" <ed********@magicinterface.com> wrote in message
news:4b**************************@posting.google.c om...

"x" <x-*****@yahoo.com> wrote in message news:<40******@post.usenet.com>...

"Daniel Morgan" <da******@x.washington.edu> wrote in message
news:1089057907.319050@yasure...

Here's what we have found: I encourage others to do so too and and in case anyone wonders who this spam-troll really is ... Domain Name:OMLET.ORG
Created On:15-Jun-2004 10:54:38 UTC ... (been in business only 2 weeks) Sponsoring Registrar:R52-LROR
Registrant Name:Amjad Daoud
Registrant Street1:Queen Noor Br 07
Registrant City:Amman
Registrant State/Province:Amman
Registrant Postal Code:9626
Registrant Country:JO ... (this is the country of Jordan)
Registrant Phone:[phone]
Registrant Email:te******************@yahoo.com Yes our troll lives in the country of Jordan. You can purchase a
product from someone to use with your valuable Oracle database
where there is not only no company behind it there is just one
guy in a third-world country.

You need to be more careful about this kind of language.
One of Amjad Daoud signatures I found on the web is: Amjad Daoud
The OMLET Team Lead
Tera Knowledge Systems, Inc.
Arlington, Texas <--------------- This is not Jordan Besides: What do you have against "guys from third-world countries" ?

He doesn't. Ok. My mistake. English is not my native tongue. :-)
Daniel, I and others have just been dealing with this
troll in the ORACLE groups for a while now.
I didn't knew that.
And you should note that
you presented the company address, which may be only a sales office
and not the location of the "OMLET Team". It certainly is different
from the residence of this guy.
I don't know if there is a company or a sale office there.
But the address is from 2001 and looks like a company address from U.S.A.
If one does business with a U.S.A. company, I think the U.S.A. and
international laws apply.
It is not relevant that the "OMLET Team" is just one guy from the country of
Jordan, regardless if this country is a "third world country" or not.
Calling Amjad a troll is certainly MUCH more polite than the crap he
has spewed in the ORACLE groups.

Probably he has done that when provoked.
Calling Jordan a third world country has nothing to do with what Amjad has
done.


-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
*** Usenet.com - The #1 Usenet Newsgroup Service on The Planet! ***
http://www.usenet.com
Unlimited Download - 19 Seperate Servers - 90,000 groups - Uncensored
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=

[x]

**** Post for FREE via your newsreader at post.usenet.com ****
"Joel Garry" <jo********@home.com> wrote in message
news:91**************************@posting.google.c om...

But check out eggy-weggies email, it's Jordanian. Not that that means
anything, I certainly am not in Tonga or oz as some of my web presence
may imply. But he says Sunnyvale, and Texas, and the registration is
on Queen Noor's... something. All-in-all, not a situation a rational
person would want to take advice or a product from. And wassup with
the 619 area code on the Sunnyvale provider, a cell?
Exactly. It doesn't mean anything.
If you want to pay him, then you may ask for a company address and
warranties.

Besides: What do you have against "guys from third-world countries" ?

There are matters of security, economics, legal liability and trust
involved.
This matters are involved anyway.
But besides that, rotten egg seems to have gone well into
net-psychoses, cyberstalking anyone who he perceives as attacking him,
which is pretty much anyone acknowledging his existence. Not a pretty
sight. DM has once again made a fast, correct call about a spammer.
You could show him some compassion then.
As far as celko, I thought DM was suitably polite in explaining the
plug issue. It was a minor gaffe on celko's part, perfectly
understandable, DM didn't "break his balls." I would say DM should
have ignored it giving the crossposting, but he clearly is making an
effort to be fair and non-partisan when netcopping. My bias is
towards letting people plug their work amid a useful answer, but DM
has almost convinced me his way is better.

Every newsgroup has his trolls :-)

-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
*** Usenet.com - The #1 Usenet Newsgroup Service on The Planet! ***
http://www.usenet.com
Unlimited Download - 19 Seperate Servers - 90,000 groups - Uncensored
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=

[Mikito Harakiri]

"Marshall Spight" <ms*****@dnai.com> wrote in message
news:8KIGc.34726$a24.6231@attbi_s03...

Methodology:

Click on the link provided by Mr. Celko.
Scroll down to the "product details" section.
Read the sales rank there (note that it's changed since then.)
--> "Amazon.com sales rank: 16,754"
subtract one :-)
--> 16,753
Go to www.google.com
search exactly as follows:
site:amazon.com "sales rank: 16753"
Click on the link

Often when a site doesn't provide some bit of
info directly, there's a way to get to it with
judicious use of google features.

For example, if you want to know if a given
32 bit number is prime, often times the fastest
way to decide is to google for it. If it's prime,
it'll be on some indexed page of prime numbers
somewhere.

Very clever. Now, I wouldn't be surprised if you invent how to do
aggregation and joins via google.

[Joel Garry]

"x" <x-*****@yahoo.com> wrote in message news:<40********@post.usenet.com>...

**** Post for FREE via your newsreader at post.usenet.com ****
"Joel Garry" <jo********@home.com> wrote in message
news:91**************************@posting.google.c om...

But check out eggy-weggies email, it's Jordanian. Not that that means
anything, I certainly am not in Tonga or oz as some of my web presence
may imply. But he says Sunnyvale, and Texas, and the registration is
on Queen Noor's... something. All-in-all, not a situation a rational
person would want to take advice or a product from. And wassup with
the 619 area code on the Sunnyvale provider, a cell?
Exactly. It doesn't mean anything.
If you want to pay him, then you may ask for a company address and
warranties.

What I want is for him to stop spamming this group. How might that
happen?

Besides: What do you have against "guys from third-world countries" ?

There are matters of security, economics, legal liability and trust
involved.

This matters are involved anyway.

They are more involved when third-world countries are mixed in. I
deal with many issues daily, living near one. There is a huge
cultural issue, too, especially having to do with dealing with
government regulations and legalities.

But besides that, rotten egg seems to have gone well into
net-psychoses, cyberstalking anyone who he perceives as attacking him,
which is pretty much anyone acknowledging his existence. Not a pretty
sight. DM has once again made a fast, correct call about a spammer.
You could show him some compassion then.

My wife's job is to show compassion. But she doesn't post on usenet.
In any unmoderated usenet group, there needs to be feedback and some
consensus about what is appropriate for the group. This group has
long taken a fairly strict no-spam stance, including creating an
appropriate group for marketing. The reason for this stance comes
from the experience of longtime posters who have seen groups get wiped
out from not taking such a strict stance. Another aspect of this
group is that it takes a critical view of postings - this is a good
thing, dba work must be precise and correct or it is bad dba work.

I think more compassion could be shown newbies, directing them towards
proper research and docs, as opposed to harsh criticism, but that's
just me, and I try to convince others of that by example rather than
argument. Unfortunately, it doesn't seem to work as well as DM's way
with newbie spammers. net.kooks do not need compassion, in fact, that
may be just the wrong thing to do, since any iota of possibility that
what they are doing isn't wrong will be rationalized into
justification.

As far as celko, I thought DM was suitably polite in explaining the
plug issue. It was a minor gaffe on celko's part, perfectly
understandable, DM didn't "break his balls." I would say DM should
have ignored it giving the crossposting, but he clearly is making an
effort to be fair and non-partisan when netcopping. My bias is
towards letting people plug their work amid a useful answer, but DM
has almost convinced me his way is better.

Every newsgroup has his trolls :-)

I don't think DM has any trolls eggy-weggy, or x-false, or whomever
you might be today.

jg
--
@home.com is bogus.
"Eggy-wegs! I'd like to crush 'em!" - Clockwork Orange