Database record format that encourages optimal retrieve(read, search, etc) & store(write, sort, etc)

    I am researching data that is optimal in how many     distinct read operations are performed over a period     of time and how many distinct write operations are performed     over a period of time.

    I have one example(bank account) which for generality is text only:

        Name: John Smith

        Age: 50

        Amount: + 1000

        Name: John Smith Jr

        Age: 31

        Amount: +100

    In the abobe example, an optimal write to the second record     would consider
    the redundancy in the Name and Amount.

    Likewise, while retrieving the record, an optimal read     will not read all the bytes of the second record rather     it will retrive the "Jr", & subtract a 0 from the Amount.

    Lest readers of this Post be confused, I admit that my     example does acknowledge the influence of      Information
    theory.(not compression, though).

    In the example I intend to have a "Live Database" that serves the purpose     of a reference point. This Live Database will have a lot     of generality(or as much as theory will permit). It is     against this reference that updates and retrieval of     the actual databases will be done.

    The above paragraph is the only use of information theory and     thereby the only limitation for my database theory.

    My questions that are relevant to database theory are:

        i) What are good formats for data that encourage Optimal read
           and optimal writes ? Sacling from intra-records(1 bit) to
           intra-record(a few bytes) to entire databases.
      

        ii) Does database theory *have to* depend upon information theory or
            parallelism or other outside domain technologies for
            efficiency ?

        iii) In my example, what could the "Live Database"  contain ?

Sincerely,
Seima Rao.                        Received on Sun Sep 13 2015 - 10:23:43 CEST