Normalisation (1 Viewer)

[tunawalker]

Hi, Just need a few pointers on how to normalise a database to reduce data redundancy.

I need to draw "tables" so as to "minimise redundant data". Does this mean I sort of draw a schema and show links and define primary and secondary keys.

How would I be best doing it?

Thanks

 

[ugly14]

well..what i would do is draw 2 tables with the fields divided in order to minimise redundant data...but i dont write any records or anything unless they ask u for it (this is rare though), just include the field names and if they're primary, secondary or foreign keys.

i dont know...but this is what i did in my trials and i got it right.

 

[seremify007]

Usually they show you an existing database right...? And the quesiton asks you to draw the new one, or if it's only a 1marker, you just identify the fields which are redundant.

I agree with what ugly14 said- there's no need to populate the database unless it's either a big question or a scenario one, in which case diagrams with data will be easier to understand.

EDIT: Not sure if I helped the TC actually... basically when identifying redundant data, I always look for things which are repeated unnecessarily (eg. Field="Colour", data="Green" "Blue", "Blue", "Blue", "Red") which could be replaced with a many to one relationship with another database whereby a colour corresponded with a number which was inserted into the Colour field. Of course in the real world, if this was a survey, there'd be idiots (like me) who would say some really obscure colour name which would thus require a more free-response field (ie. text), but then... this is IPT