search

fasterthanlime / cs322

CS-323 project
2 stars 1 forks source link

ER model for the data #1

Closed greut closed 12 years ago

greut commented 12 years ago

Creating the EF model from the data.

I personaly uses Dia which is crossplatform, so will produce a lovely png (svg) along the .dia file.

nddrylliog commented 12 years ago

So, we don't have an ER model yet, but! We have studied the sample data, and came up with tables+fields that somewhat make sense. Deriving an ER model from there should be easy.

We did that on paper, so I'll upload a picture of that soon.

greut commented 12 years ago

Heya, can you give a look at the ER model? I've tried to explain some choices into design.md (in case). What's next?

greut commented 12 years ago

Aren't first_season/last_season denormalized fields too?

nddrylliog commented 12 years ago

What do you mean by denormalized fields?

greut commented 12 years ago

Data that duplicates information you have but are too laty to recalculate it everytime, like a blo post comments counter. I'll ask the TA about those and see. Performance is the root of all evil :-)

nddrylliog commented 12 years ago

Oh. Yeah I see exactly what you mean now. I didn't think it was in the scope of the project to 'clean up the dataset' and remove denormalized data. We did already (with Sebastian) think of way to reduce duplication of data (with the 'human' table for example) thought this was a reasonable compromise.

Curious to see the outcome of the TA discussion though!

SeZuo commented 12 years ago

Cardinality (1..n) should not be written in the ER model. That info is given by the arrows and the bold lines. However I think it should be added to the logical design.

Also the weak entities should have the relation AND the entity in bold.

Tell me if I'm wrong.

greut commented 12 years ago

How do you differenciate between 0..n and 1..n ? We can put everything all those in the logical design, if you want to take over @nddrylliog on that, be my guest mate.

nddrylliog commented 12 years ago

@greut 0..n is a simple line, 1..n is a bolded line?

I don't think it belongs in the logical model, at least from the example which I reproduced they didn't have that kind of details, it's just foreign keys etc.

SeZuo commented 12 years ago

@nddrylliog I'm not sure for the logical.

@greut for the ER you use a normal line for 0..n, a bold line for 1..n, an arrow for 0..1 and a bold arrow for 1..1.

SeZuo commented 12 years ago

For the logical, I've checked the lecture and they never use any kind of drawing, only the SQL queries. I suppose we can use any convention we want then.

greut commented 12 years ago

So you didn't like the “isa” on the stats? It's really what it is.

SeZuo commented 12 years ago

ISA would be fine. I've only tried to add constraints by removing it.

Another thought: aren't TEAM_STATS the sum of all the player stats playing in the team? (For the first assignment we should avoid to over-simplificate, I leave this for assignment 2).

SeZuo commented 12 years ago

Comments on the last commit:

Orthography

Discussion