|
Home > Archive > microsoft.public.cert.exam.mcsd > July 2002 > Re: E-R Diagrams and 70-100
You are viewing an archived Text-only version of the thread.
To view this thread in it's original format and/or if you want to reply to
this thread please [click here]
| Author |
Re: E-R Diagrams and 70-100
|
|
| James Norman 2002-06-22, 7:14 pm |
| Thanks Patrick, actually I had checked that one out from the library
on Tuesday night. This enthuses me even more now to finish reading
it. I still need to see if the book Alan suggested was available.
(Understanding "real life" diagrams is a long term project, but the
more I understand before the test....)
I'm getting close to developing the methodology I need to beat
"Microsoft's Thinking" on 70-100's ER diagrams (which judging by the
transcender tests, I may or may not need, depending on the luck of the
draw)
(Here's where Steve's post really helped)
1. Entities in a real world E-R diagram are nouns. Entities on
70-100, based on what I've seen in Transcender, are tables.
2. The Transcender test seems to only offer 1-M and 1-1 for the
relationships. So instead of having a:
(X)<--- Many to Many --> (Y)
or
(X) <---- Relationship verb Z --> (Y)
relationship. You probably have a third entity (table) which can fit
between the two and be drawn like this.
(X) <--- One to Many --> (Z as a noun) <--- Many to One --> (Y)
Therefore, when you see three entities on the test that look like this
(X) (Y) (XY)
or this
(X) (Y) (XY NOUN)
or this
(X) (Y) (Z)
Then it's a probably cinch they are looking to form a 1->M->1
relationship here.
Another two observations I made were this
3. For the total count of entities that can be used (I'm not assuming
all entities in the diagram can be used) the total number of
relationships is N or N-1. And directly or indirectly, they will all
be linked. For Example
(Count of Entities = Count of Relationships)
(A) <--- relationship --> (B) <-- relationship --> (C) <--
relationship --> (A)
Which while not represented as such here, could be drawn as a circle
(or more accurately, triangle) 3 Entities, 3 relationship. All tied
together.
(Count of Entities - 1 = Count of Relationships)
(A) <--- relationship --> (D)
(A) <--- relationship --> (B) <-- relationship --> (C)
All relationships can directly (B,D) or indirectly (C) make their way
back to A.
4 entities, 3 relationships.
That sounds nuts, but we're talking test strategy here for now.
Though I'm wondering if I can tie this theory into real life E-R
diagrams. I mean, knowing in advance how many relationships I SHOULD
have based on the number of Entities identified is a useful way for me
to self-validate this. Assuming the theory is true.
4. There seem to be intentional deceptions thrown in (again based on
the Transcender test.) For example
An (A) can have one to many (B). (B) can have exactly one (C) but (C)
can have more than one (B) An (A) can have one to many (C)
To me this would be an implied relationship such as this
(A) <-- One to Many --> (B) <-- Many to One --> (C)
In other words, another implied MANY TO MANY relationship.
But instead on the practice tests the right answer is
(A) <-- One to Many --> (C)
(B) <-- Many to One --> (C)
I still don't get *why* that's right, but I noticed the test text
emphasized the relation of A and B to C. The *why* is important to
me, and I still need to figure it out.
This last one is important because consider the point difference. Two
wrong answers -8 (though later reweighted..I'm betting to -30 actual
exam points) Two right answers +8. I think more than anything else
this is the one thing that tripped me up.
Any thoughts, or am I just crazy? (I know, not an "or" question.)
"Patrick" <patrickNOJUNK@pacific.demon.co.uk> wrote in message news:<3_0J8.100$Wr2.68262@news.lhr.globix.net>...
> Another book that may be of interest to you is "Database Design for Mere
> Mortals", which I used to give me a decent grounding in database design.
>
> Link to Amazon (US) for more details:
>
> http://www.amazon.com/exec/obidos/A...655/sr=1-1/ref=
> sr_1_1/104-3619280-0215959
>
> (watch for wrapping)
>
>
> Patrick
>
>
| |
| Patrick 2002-06-22, 7:14 pm |
| James,
To be honest, reading Steve's stuff left me with the shakes - very intense
stuff that I had trouble understanding!! But the jist of it does apply to
both the Transcenders and the real exam.
The main thing that I had trouble with doing the Transcenders was realising
that just because a relationship was 1 to 1 in one direction, it can easily
be Many to 1 in the other direction!! Sounds obvious, but when you're
focusing on a particular table, and it's relationships, it is easy to forget
that relationships are 2 way.
Can't remember if I posted this to you already, but I posted my views about
the exam not long after taking it here:
http://groups.google.com/groups?hl=...e=off&selm=IWht
8.18%24qo4.3937%40news.lhr.globix.net
{again, watch for the wrapping)
Anyway, let us know how you get on, and Good Luck.
Patrick
"James Norman" <jnorman@Inficad.com> wrote in message
news:878f46a4.0205302249.83b503a@posting.google.com...
> Thanks Patrick, actually I had checked that one out from the library
> on Tuesday night. This enthuses me even more now to finish reading
> it. I still need to see if the book Alan suggested was available.
> (Understanding "real life" diagrams is a long term project, but the
> more I understand before the test....)
>
> I'm getting close to developing the methodology I need to beat
> "Microsoft's Thinking" on 70-100's ER diagrams (which judging by the
> transcender tests, I may or may not need, depending on the luck of the
> draw)
>
> (Here's where Steve's post really helped)
>
> 1. Entities in a real world E-R diagram are nouns. Entities on
> 70-100, based on what I've seen in Transcender, are tables.
>
> 2. The Transcender test seems to only offer 1-M and 1-1 for the
> relationships. So instead of having a:
>
> (X)<--- Many to Many --> (Y)
> or
> (X) <---- Relationship verb Z --> (Y)
>
> relationship. You probably have a third entity (table) which can fit
> between the two and be drawn like this.
>
> (X) <--- One to Many --> (Z as a noun) <--- Many to One --> (Y)
<snip really in depth stuff>
> 4. There seem to be intentional deceptions thrown in (again based on
> the Transcender test.) For example
>
> An (A) can have one to many (B). (B) can have exactly one (C) but (C)
> can have more than one (B) An (A) can have one to many (C)
>
> To me this would be an implied relationship such as this
>
> (A) <-- One to Many --> (B) <-- Many to One --> (C)
>
> In other words, another implied MANY TO MANY relationship.
>
> But instead on the practice tests the right answer is
>
> (A) <-- One to Many --> (C)
> (B) <-- Many to One --> (C)
>
> I still don't get *why* that's right, but I noticed the test text
> emphasized the relation of A and B to C. The *why* is important to
> me, and I still need to figure it out.
>
> This last one is important because consider the point difference. Two
> wrong answers -8 (though later reweighted..I'm betting to -30 actual
> exam points) Two right answers +8. I think more than anything else
> this is the one thing that tripped me up.
>
>
> Any thoughts, or am I just crazy? (I know, not an "or" question.)
>
| |
| Alan Davis 2002-06-22, 7:14 pm |
| On 30 May 2002 23:49:11 -0700, jnorman@Inficad.com (James Norman)
wrote:
>Thanks Patrick, actually I had checked that one out from the library
>on Tuesday night. This enthuses me even more now to finish reading
>it. I still need to see if the book Alan suggested was available.
>(Understanding "real life" diagrams is a long term project, but the
>more I understand before the test....)
>
>I'm getting close to developing the methodology I need to beat
>"Microsoft's Thinking" on 70-100's ER diagrams (which judging by the
>transcender tests, I may or may not need, depending on the luck of the
>draw)
>
>(Here's where Steve's post really helped)
>
>1. Entities in a real world E-R diagram are nouns.
Not all 'nouns' identified will become entity. It may be that all the
information required for that 'noun' could be described in a single
attribute.
Also in identifying a 'noun' may also lead to identifying super
classes of that 'noun' which may form a 1:1 'isa' type relationship.
an example would an 'Employee' (super class) 'is an ' Hourly Paid
Staff' or is a 'Salary Staff'.
>Entities on 70-100, based on what I've seen in Transcender, are tables.
>
>2. The Transcender test seems to only offer 1-M and 1-1 for the
>relationships. So instead of having a:
>
> (X)<--- Many to Many --> (Y)
>or
> (X) <---- Relationship verb Z --> (Y)
>
>relationship. You probably have a third entity (table) which can fit
>between the two and be drawn like this.
>
> (X) <--- One to Many --> (Z as a noun) <--- Many to One --> (Y)
>
>Therefore, when you see three entities on the test that look like this
>
>(X) (Y) (XY)
>
>or this
>
>(X) (Y) (XY NOUN)
>
>or this
>(X) (Y) (Z)
>
>Then it's a probably cinch they are looking to form a 1->M->1
>relationship here.
>
>
>Another two observations I made were this
>
>3. For the total count of entities that can be used (I'm not assuming
>all entities in the diagram can be used) the total number of
>relationships is N or N-1. And directly or indirectly, they will all
>be linked. For Example
>
>(Count of Entities = Count of Relationships)
>(A) <--- relationship --> (B) <-- relationship --> (C) <--
>relationship --> (A)
>
>Which while not represented as such here, could be drawn as a circle
>(or more accurately, triangle) 3 Entities, 3 relationship. All tied
>together.
>
>(Count of Entities - 1 = Count of Relationships)
Not necessarily. you could have any number of recursive relationships,
i.e. an Employee 'is managed' by another Employee. In which case the
Employee entity would contain a foreign key referencing back to
itself.
>
>(A) <--- relationship --> (D)
>(A) <--- relationship --> (B) <-- relationship --> (C)
>
>All relationships can directly (B,D) or indirectly (C) make their way
>back to A.
>
>4 entities, 3 relationships.
>
>That sounds nuts, but we're talking test strategy here for now.
>Though I'm wondering if I can tie this theory into real life E-R
>diagrams. I mean, knowing in advance how many relationships I SHOULD
>have based on the number of Entities identified is a useful way for me
>to self-validate this. Assuming the theory is true.
>
>4. There seem to be intentional deceptions thrown in (again based on
>the Transcender test.) For example
>
>An (A) can have one to many (B). (B) can have exactly one (C) but (C)
>can have more than one (B) An (A) can have one to many (C)
>
In relational theory this is defined as the 'degree of a relationship'
(1:n, 1:1 etc) and the 'participation' (optional or mandatory)
>To me this would be an implied relationship such as this
>
>(A) <-- One to Many --> (B) <-- Many to One --> (C)
>
>In other words, another implied MANY TO MANY relationship.
>
>But instead on the practice tests the right answer is
>
>(A) <-- One to Many --> (C)
>(B) <-- Many to One --> (C)
>
>I still don't get *why* that's right, but I noticed the test text
>emphasized the relation of A and B to C. The *why* is important to
>me, and I still need to figure it out.
>
>This last one is important because consider the point difference. Two
>wrong answers -8 (though later reweighted..I'm betting to -30 actual
>exam points) Two right answers +8. I think more than anything else
>this is the one thing that tripped me up.
>
>
>Any thoughts, or am I just crazy? (I know, not an "or" question.)
>
From an E-R diagram you will build a rotational model. As you have
already worked out, you cannot implement a m:n relationship, thus
typically an 'intersection relation' is formed. When taking about a
relational model, 'entities' become know as 'Relations' and can get
somewhat confusing with what seem like changes in terminology.
Relation database theory can get very heavy and there is far more to
it than E-R diagrams. If get hold of that book I mentioned, you with
understand what I have failed to explain!
| |
| James Norman 2002-06-22, 7:14 pm |
| I'd like to thank everyone in this thread for their help. I'm glad I
asked the question, because even though I didn't encounter a single
E-R diagram on my exam ($@%^!#!) it's still useful in real life.
This is the first MS Exam in a long while where:
1. I felt I passed by a wide margin, just to have it be by a narrow
margin (772)
2. My actual score was SIGNIFICANTLY different than the Transcender
score (though when I took Exam C for the first time the other day, I
think the difference was ONE POINT. Today it was closer to 100 and I
made a point of not looking at Exam C's answers.)
Got lots of "Build a tree" and "reorder questions." I also got a DFD
(which thankfully, I'm much better at than E-R.)
I don't know what other thoughts I can offer here. My mind is
scrambled from being up almost 20 straight hours now.
| |
| DavidW 2002-07-20, 12:25 am |
| I haven't done the 70-100 yet - (I've done 70-176/176 and I'm working on
229) - but it's coming soon....
So, are you (all) saying that, basically, you get a lot of questions with
pictures of tables in a database, and you're supposed to connect PK-Fk
relationships????
(just trying to get a handle on what's required on the test)
David Wier
http://aspexpress.com
"James Norman" <jnorman@Inficad.com> wrote in message
news:878f46a4.0206031115.11c8af91@posting.google.com...
> I'd like to thank everyone in this thread for their help. I'm glad I
> asked the question, because even though I didn't encounter a single
> E-R diagram on my exam ($@%^!#!) it's still useful in real life.
>
> This is the first MS Exam in a long while where:
>
> 1. I felt I passed by a wide margin, just to have it be by a narrow
> margin (772)
> 2. My actual score was SIGNIFICANTLY different than the Transcender
> score (though when I took Exam C for the first time the other day, I
> think the difference was ONE POINT. Today it was closer to 100 and I
> made a point of not looking at Exam C's answers.)
>
> Got lots of "Build a tree" and "reorder questions." I also got a DFD
> (which thankfully, I'm much better at than E-R.)
>
> I don't know what other thoughts I can offer here. My mind is
> scrambled from being up almost 20 straight hours now.
| |
|
| It's even more abstract than that. First, it's more of a diagram tool. The
picture of tables is not already made, you have to choose which tables to
put into the diagram - not all available in some cases. Second, rather than
pk/fk connections, you then create one to many or one to one connections
between the entities.
MS has a download available with a little demo of some similar questions. It
used to be available from the page for the 70-100 exam (and probably the
other design exams).
The same format question has also been used to have you create a data flow
diagram - with each box representing a system or process, and the
connections represent different types or states of data.
"DavidW" <dwier@davidwier.com> wrote in message
news:#sVKcL7LCHA.2452@tkmsftngp08...
> So, are you (all) saying that, basically, you get a lot of questions with
> pictures of tables in a database, and you're supposed to connect PK-Fk
> relationships????
| |
|
|
| Patrick 2002-07-22, 4:25 am |
| It might be worth downloading the Transcender (free) demo for this exam just
to get a better feel for the question format.
--
Patrick Mullin,
MCSD (VB6).
Email <first name> at AuroraDigitalis dot com
"Tim" <spammers@begone> wrote in message
news:ufp41z$LCHA.2424@tkmsftngp09...
> It's even more abstract than that. First, it's more of a diagram tool. The
> picture of tables is not already made, you have to choose which tables to
> put into the diagram - not all available in some cases. Second, rather
than
> pk/fk connections, you then create one to many or one to one connections
> between the entities.
>
> MS has a download available with a little demo of some similar questions.
It
> used to be available from the page for the 70-100 exam (and probably the
> other design exams).
>
> The same format question has also been used to have you create a data flow
> diagram - with each box representing a system or process, and the
> connections represent different types or states of data.
>
> "DavidW" <dwier@davidwier.com> wrote in message
> news:#sVKcL7LCHA.2452@tkmsftngp08...
>
> > So, are you (all) saying that, basically, you get a lot of questions
with
> > pictures of tables in a database, and you're supposed to connect PK-Fk
> > relationships????
>
>
>
|
|
|
|
|