Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

SUO: Re: Relations And Their Divisitudes




o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o

RATD.  Note 8

o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o

Tom,

Shifting to another thread where I have
all of my asscoiative facilities intact.

I suggested that we begin by checking whether we have
the same concepts of 1-adic and 2-adic projections as
applied to 3-adic relations.

As I read your table I see a possible divergence on this row:

o-----------------------------o---------------------------------------o
| 1-adic projection           | result of a relational PROJECT        |
|                             | operation on a table, that            |
|                             | leaves just one column                |
o-----------------------------o---------------------------------------o

We are talking about unkeyed tables, right?
Thus, when we extract a column from a table,
only the set of entries counts, not the set
with multiplicites, what some call a "bag",
and certainly not an ordered list of items,
which would amount to a column plus the key.

In other words, I am defining the 1-adic and 2-adic
projections on 3-adic relations in the following way:

The 1-adic "projections" Proj_X, Proj_Y, Proj_Z,
alternatively written as p_1,    p_2,    p_3,
as applying to a 3-adic relation L c X x Y x Z,
along with the equivalent forms of application
L_X = p_1 (L), L_Y = p_2 (L), L_Z = p_3(L),
respectively, are defined as follows:

Proj_X (L)  =  L_X  =  {x in X : <x, y, z> in L for some y in Y, z in Z},

Proj_Y (L)  =  L_Y  =  {y in Y : <x, y, z> in L for some x in X, z in Z},

Proj_Z (L)  =  L_Z  =  {z in Z : <x, y, z> in L for some x in X, y in Y}.

The 2-adic "projections" Proj_XY, Proj_XZ, Proj_YZ,
alternatively written as p_12,    p_13,    p_23,
as applying to a 3-adic relation L c X x Y x Z,
along with the equivalent forms of application
L_XY = p_12(L), L_XZ = p_13(L), L_YZ = p_23(L),
respectively, are defined as follows:

Proj_XY (L)  =  L_XY  =  {<x, y> in XxY : <x, y, z> in L for some z in Z},

Proj_XZ (L)  =  L_XZ  =  {<x, z> in XxZ : <x, y, z> in L for some y in Y},

Proj_YZ (L)  =  L_YZ  =  {<y, z> in YxZ : <x, y, z> in L for some x in X}.

Well, it's late, and the risk of typos and worse grows with
this sort of stuff, so I will leave it there for the moment.

Jon Awbrey

o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o