[Discuss] SQL discussion

[markw at mohawksoft.com (markw at mohawksoft.com)]

> You can and usually have many tables.

That's true.

This example is quite good, actually.

So the issue with dimensions, regardless of technology, is how do you
represent them. A computer, at its core, only has one axis, i.e. element
offset linear addressing. Everything else is a construct beyond that. For
example:

mov ecx, index
mov ebx, [ecx]

Everything else is X * Y [Z [* n]]
Where X is the width of the record or data item, and x is the index.

If you want to represent multiple dimensions you need to come up with a
way to map the above into some addressing scheme that emulates true
n-dimensions.

What ever system you use to access a multidimensional array it has to
construct this behavior in what ever language it is written in, and what
ever functions and constructs it uses to implement these arrays, they are
still constructs based on linear addressing.

SQL should is no better or worse than any other system that is capable of
multidimensional representation. Probably better because people have been
doing it in SQL for some time.

>
> On 01/14/2015 04:22 PM, Mike Small wrote:
>> Richard Pieri <richard.pieri at gmail.com> writes:
>>> Precisely. What is the structure of a relational database? A table. A
>>> 2-dimensional table. If you have 3 dimensions of data in a relational
>> |   x |   y |  z |     t | Humidity | Pressure |
>> |-----+-----+----+-------+----------+----------|
>> | 100 | -10 | 12 | 12:05 |       40 |    1.302 |
>> ...
>> |-----+-----+----+-------+----------+----------|
>>
>>
>> Not what you had in mind?
>>
>>
>
> --
> Jerry Feldman <gaf at blu.org>
> Boston Linux and Unix
> PGP key id:B7F14F2F
> PGP Key fingerprint: D937 A424 4836 E052 2E1B  8DC6 24D7 000F B7F1 4F2F
>
>
> _______________________________________________
> Discuss mailing list
> Discuss at blu.org
> http://lists.blu.org/mailman/listinfo/discuss
>