Normalization – 1NF, 2NF, 3NF and 4NF




Database Normal Forms.

Original source


31 responses to “Normalization – 1NF, 2NF, 3NF and 4NF”

  1. For the newsletter it seems that this solution will consume memory over time. We can have a bitmap or Boolean columns to represent if the client is subscribed or not for the newsletter since the company has only two newsletters meanwhile we can also expand the model later to distinguished table with respect to the previous formal norms into a separate table if the company has more newsletters. And I highly doubt that any company has more than 10 newsletters except for marketing companies that have a complex newsletter structure in behalf of their clients.

    Technically speaking:

    A Boolean can consume 1 bit of information, in this case it will consume in the worst scenario 2n bits – where n is the number of records i.e. rows.
    A character will consume one byte (8 bits) which means in the worst scenario they will all subscribe to both of the newsletters meaning 24*8*n bits (counting space as a character)
    We all agree here that 2*n < 24*8*n -> 2*n < 3(8*8)*n -> 2*n < 3(2^3*2^3)*n -> 2*n <<< 2^6*n this last result shows how bad it can get when the rows number goes upper since they are both of the same degree I think you can plot it in you mind and imagine the memory disaster on a long term plan.

Leave a Reply