1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
|
********************************************************************************
Changes that were made in July 2006 by Joshua Reich I.
********************************************************************************
Code Cleanup:
Update the calling convention for all external facing functions. By external
facing, I mean all functions that are directly referenced in cube.sql. Prior
to my update, all functions used the older V0 calling convention. They now
use V1.
New Functions:
cube(float[]), which makes a zero volume cube from a float array
cube(float[], float[]), which allows the user to create a cube from
two float arrays; one for the upper right and one for the lower left
coordinate.
cube_subset(cube, int4[]), to allow you to reorder or choose a subset of
dimensions from a cube, using index values specified in the array.
********************************************************************************
Changes that were made in August/September 2002 by Bruno Wolff III.
********************************************************************************
Note that this was based on a 7.3 development version and changes may not
directly work with earlier versions.
I fixed a bug in cubescan.pl that prevented signed numbers with no digits
before a decimal point from being accepted. This was submitted as a separate
patch and may already be applied.
cube_inter should really return NULL if the two cubes don't overlap. However
this requires changing to the new calling sequence and I don't know enough
about how to do it to make the change.
I changed all floats to doubles except for g_cube_penalty which I don't
think can be changed to return double. This might cause the penalty to
overflow sooner than one might expect, but overflow could have happened
even with floats.
I changed the output format (in cube_out) to use %.16g instead of %g, since the
default is only 6 digits of precision.
I changed all of the functions declared with (isstrict) to use the current
method of declaring this.
I changed all of the externally visible functions to be immutable which
they are. I don't think this matters for the gist functions and didn't
try to declare them immutable in case there was something tricky about them
that I don't understand.
I changed the regression tests to use some larger exponents to test output
in exponential form. 1e7 was too small for this.
I added some regression tests to check for 16 digits of precision. This
may or may not be a good idea.
I got rid of the swap_corners function. It created scratch boxes that
were iterated through and deleted. This is slower than just getting the
larger or smaller coordinate as needed, since swap_corners was doing the
same thing with the overhead of a function call and memory allocation.
I added memset calls to zero out newly allocated NDBOXes as the documentation
on functions indicates should be done. This still doesn't allow a hash
index for equality since there are multiple representations of the
same cube.
I got rid of a call to cube_same in cube_lt and cube_gt since the test
was redundant with other checks being made. The call to cube_same would
only be faster if most of the time you were comparing equivalent cubes.
In cube_lt and cube_gt, the second (UR) for loop for comparing
extra coordinates to 0 had the wrong range.
Note that the cube_distance function wasn't mentioned in the README.cube file.
I added regression tests for the cube_distance function.
I added the following new functions:
cube
cube_dim
cube_ll_coord
cube_ur_coord
cube_is_point
cube_enlarge
cube takes text input and returns a cube. This is useful for making cubes
from computed strings.
cube_dim returns the number of dimensions stored in the data structure
for a cube. This is useful for constraints on the dimensions of a cube.
cube_ll_coord returns the nth coordinate value for the lower left corner
of a cube. This is useful for doing coordinate transformations.
cube_ur_coord returns the nth coordinate value for the upper right corner
of a cube. This is useful for doing coordinate transformations.
cube_is_point returns true if a cube is also a point. This is true when the
two defining corners are the same.
cube_enlarge increases the size of a cube by a specified radius in at least
n dimensions. If the radius is negative the box is shrunk instead. This
is useful for creating bounding boxes around a point for searching for
nearby points. All defined dimensions are changed by the radius. If n
is greater than the number of defined dimensions and the cube is being
increased (r >= 0) then 0 is used as the base for the extra coordinates.
LL coordinates are decreased by r and UR coordinates are increased by r. If a
LL coordinate is increased to larger than the corresponding UR coordinate
(this can only happen when r < 0) than both coordinates are set to their
average.
I added regression tests for the new functions.
I added documentation for cube_distance and the new functions to README.cube
as well as making a few other minor changes.
I changed create function to create or replace function in the install
script.
I limited the number of dimensions allowed in cube_enlarge and cube_in
to 100 to make it harder for people to mess up the database. The constant
is defined in cubedata.h and can be increased if you need something larger.
I added grant statements to the install script to make the functions
executable to everyone.
Bruno Wolff III <bruno@wolff.to>
|
