Hi Steve
1) I may have 10 to 20 rows of data with phrases. I count the number
of rows then make the Random function statement to pick one of
the rows / phrases at random. Each user can have as many or as
little rows of data as they like.
2) It seems each time I restarted Rbase it would pick the same sequence
of numbers such as 0 0 5 0 5 3 8 2 10 14 , then restart and the same
Random statement would give the same numbers. If I did not close
Rbase then the numbers kept going and seemed to be random.
One thing I did not test is we get in and out of this form and have
different
categories of phrases, so the Random statement may be set to 5 or 10 or
20
for any given office depending on the category they are trying to get a
random
phrase for. So, since the Random statement is not exactly the same the
numbers
may not be very random. I really have not checked this yet.
I do not need 100% random numbers and a few repeats is OK. But if I
use this
form 20 times a day and sometimes I am doing Random for x rows then
Random for
y rows, then Random for x .... I would not want the same set of numbers
like the
above example 0 0 5 0 5 3 8 2 10 14 . Because I would keep geting 0
each time that
Randon for x was run.
I hope that makes sense, again I need to test it further with this new
variable included.
Thanks
Marc
Marc, couple of Q's:
(1) How many of these text items/phrases exist and are they identified
by some sort of numeric?
(2) I still don't understand your perspective that "... RBase ...
remembers where it left off ...". Could you elaborate on that?
Now, let me try somethin' out on you. Let's say that you have 15 items
- I take that from your RANDOM(15) example - and that they are
identified numerically from 1 thru 15. If so, then perhaps something
along the lines of the command-file below would meet your needs. Again,
I cannot say, Y||N, if the results are "precisely" random; of course, I
cannot precisely define "random", either, at least in a rock-solid
statistical/mathematical sense. I do know that there exists one-or-more
statistical tests to verify something like this.
However, to my eyeball-test, it appears random, despite the occasional
sets of repeated values which might give one pause. Still, no number
repeats more than twice. Were we to loop it 1000x rather than 100,
then, I think, our odds of getting 3 of the same values in a row should
increase, but, to be honest, I don't know if I could have determined by
how much even when I was studying business/management statistics.
However, I do know that the probability of a series of independent
events is pretty simply calculated by multiplying the probability of a
single event by the number in the series, i.e. P(A1,A2,A3) = P(A1) *
P(A2) * P(A3), where A is the generation of that random number and 1,2,3
are iterations 1..3. In our case, RBase tells us:
R>SET VAR vProbability03 REAL=((1/15) * (1/15) * (1/15))
R>sho var
Variable = Value Type
------------------ ------------------------------ --------
#DATE = 04/09/2008 DATE
#TIME = 10:41:14.695 TIME
#PI = 3.14159265358979 DOUBLE
SQLCODE = 0 INTEGER
SQLSTATE = 42S02 TEXT
#NOW = 04/09/2008 10:41:14.787 DATETIME
vProbability03 = 0.0002963 REAL
I do think that probability goes up with the number of iterations.
However, I think it remains sufficiently low. Again, that's beyond what
I currently grasp. This site, http://mathforum.org/dr.math/ask/, has
some pretty good stuff, if anybody's interested, and I'm sure there are
more.
The point of all this is, that I ain't so sure that RB's RANDOM()
function has any faults. I really think it works as advertised. I
think it comes down to how we utilize it, including what we pass to it
and what we do with the value returned from it. Albert's example is a
nice variation on doing something with what we pass to it. I hope my
use of MOD(.vAlbertsNUM,IntegerOfPossibleOutcomes)
illustrates a variation on doing something with the return value.
Now, let me ask/add, if any repeating series is "verboten", i.e. no
number can occur more than one at a time, then you could add another
WHILE-loop to trap for this, as shown next, between the long
asterisked-lines. This block will loop anytime the number,
vWhichTextToGet, repeats. So, there will never be a series of numbers
longer than 1. Truly, although this approach seems to nicely pass the
"eyeball test", I suspect that, mathematically/scientifically, it's
actually "less" random ...
-- ****************************
-- *** RandomTextTestxA.RMD ***
-- ****************************
...
-- *** Init var to help trap repeaters ... ***
SET VAR vWhichTextToGetPREV INT = 0
...
WHILE ...
-- ******************************************************
-- *** Is the current value the same as the previous? ***
-- *** This will always be true the first iteration ***
-- *** as they are both zero ***
WHILE vWhichTextToGetPREV = .vWhichTextToGet THEN
-- *** Albert's expression ; check your quote settings ...***
SET VAR vAlbertsNUM INT=(RANDOM(INT(1000 *
(FLOAT(FORMAT(.#time,'hhmmss.sss'))))))
-- *** Assign value to determine which item to select from table
************
-- *** 15 is the value because it is the count of our RandomText
records ***
-- *** Since result of MOD() can be 0..14, add one to make it
1..15, ***
-- *** as 15 is our "sample space", possible outcomes, or solution
domain ***
SET VAR vWhichTextToGet = ((MOD(.vAlbertsNUM,15)) + 1)
ENDWHILE
-- *** Assign value to duplicate test var ***
SET VAR vWhichTextToGetPREV = .vWhichTextToGet
-- ******************************************************
...
ENDWHILE
...
RETURN
Well, not to beat a dead horse, but "Hi, ho, Silver, away!"
Steve in Memphis
-- ****************************
-- *** RandomTextTestxA.RMD ***
-- ****************************
SET VAR vWhichTextToGet INT=0
SET VAR vCounter INT=0
WHILE vCounter < 100 THEN
-- *** Albert's expression ; check your quote settings ...***
SET VAR vAlbertsNUM INT=(RANDOM(INT(1000 *
(FLOAT(FORMAT(.#time,'hhmmss.sss'))))))
-- *** Assign value to determine which item to select from table
************
-- *** 15 is the value because it is the count of our RandomText
records ***
-- *** Since result of MOD() can be 0..14, add one to make it 1..15,
***
-- *** as 15 is our "sample space", possible outcomes, or solution
domain ***
SET VAR vWhichTextToGet = ((MOD(.vAlbertsNUM,15)) + 1)
-- *** See what ya' got there ************************************
-- *** In my simple table, I used the numbers 1..15 for ID's ***
-- *** and the name of that number for the text, like NUM2Word ***
SELECT .vCounter, .vAlbertsNUM, .vWhichTextToGet, RandomText, +
FROM RANDOM_TEXT_TBL +
WHERE RandomTextID = .vWhichTextToGet
-- *** Just for appearance during testing, after first SELECT ***
SET HEADINGS off
-- *** Use something like this to SELECT the RandomText INTO a var
***
{
SELECT RandomText INTO vRandomText IND viRandomText +
FROM MY_RANDOM_TEXT_TBL +
WHERE RandomTextID = .vWhichTextToGet
}
-- *** Increment counter
SET VAR vCounter = (.vCounter + 1)
ENDWHILE
-- *** Reset headings
SET HEADINGS on
RETURN
R>run spRandomTextTestxA.RMD
0 5976443 9 RandomText
---------- ---------- ---------- ------------------------ 0
5976443 9 Nine
1 73764218 9 Nine
2 54736378 14 Fourteen
3 86170466 12 Twelve
4 71083348 14 Fourteen
5 9630739 5 Five
6 63371555 6 Six
7 82733176 2 Two
8 8318094 10 Ten
9 14836858 14 Fourteen
10 7375324 5 Five
11 62039506 2 Two
12 36737584 5 Five
13 22929708 4 Four
14 46234866 7 Seven
15 11494059 10 Ten
16 47519728 14 Fourteen
17 901058 9 Nine
18 25827580 11 Eleven
19 28052423 9 Nine
20 20932946 12 Twelve
21 69228632 3 Three
22 14027609 15 Fifteen
23 27557416 2 Two
24 73266756 7 Seven
25 56107688 9 Nine
26 9199715 6 Six
27 88609858 14 Fourteen
28 51146324 15 Fifteen
29 32187899 15 Fifteen
30 37580366 12 Twelve
31 83932191 7 Seven
32 3137028 4 Four
33 45993076 2 Two
34 78564800 6 Six
35 31715155 11 Eleven
36 79132163 9 Nine
37 59564663 9 Nine
38 31459319 15 Fifteen
39 2155319 15 Fifteen
40 2736561 7 Seven
41 35934060 1 One
42 73965057 13 Thirteen
43 9230337 13 Thirteen
44 38940407 3 Three
45 84433007 3 Three
46 76837942 8 Eight
47 89895040 11 Eleven
48 48220837 8 Eight
49 68672782 8 Eight
50 65232620 6 Six
51 89416762 8 Eight
52 7475500 11 Eleven
53 41259878 9 Nine
54 26720475 1 One
55 84079964 15 Fifteen
56 46963867 8 Eight
57 89914673 9 Nine
58 54044483 9 Nine
59 48696500 6 Six
60 11897410 11 Eleven
61 47798232 13 Thirteen
62 33859521 7 Seven
63 64665439 5 Five
64 60752484 10 Ten
65 22017738 4 Four
66 51872545 11 Eleven
67 40520219 15 Fifteen
68 30703042 8 Eight
69 6054396 7 Seven
70 87617733 4 Four
71 58455404 15 Fifteen
72 84725450 6 Six
73 76885613 9 Nine
74 78214981 2 Two
75 45392731 2 Two
76 6176772 13 Thirteen
77 80356453 14 Fourteen
78 17896237 8 Eight
79 64351378 14 Fourteen
80 19973710 11 Eleven
81 32282772 13 Thirteen
82 79071673 14 Fourteen
83 15754822 8 Eight
84 50507216 12 Twelve
85 47684427 13 Thirteen
86 57215205 1 One
87 44483419 5 Five
88 22028960 6 Six
89 23158085 6 Six
90 39071453 9 Nine
91 77811971 12 Twelve
92 25371841 2 Two
93 71949472 8 Eight
94 47245095 1 One
95 63364283 9 Nine
96 40589974 5 Five
97 22996820 6 Six
98 13880414 15 Fifteen
99 61303552 8 Eight
-----Original Message-----
From: [email protected] [mailto:[EMAIL PROTECTED] On Behalf Of Marc
Sent: Wednesday, April 09, 2008 1:27 AM
To: RBASE-L Mailing List
Subject: [SPAM] [RBASE-L] - Re: [SPAM] [RBASE-L] - Random numbers
Importance: Low
Thanks Albert
But I do not think that will work for me.
I use the random number to get a row number.
I have a table with dozens of sample text for form letters
and I want to select them at Random.
I think the Random function will work OK, as long as they
do not restart Rbase several times a day. It seems as long
as RBase is on it remembers where it left off so to speak
then picks another random number but is you restart RBase
it starts over with the same sequence of numbers.
So, I think this will work, I never really tested it until now.
Thanks
Marc
----- Original Message ----- From: "Albert Berry"
<[EMAIL PROTECTED]>
To: "RBASE-L Mailing List" <[email protected]>
Sent: Tuesday, April 08, 2008 12:11 PM
Subject: [RBASE-L] - Re: [SPAM] [RBASE-L] - Random numbers
R:Base uses a pseudo random number generator. You can use the system
time variable to the 1/1,000 second to generate pseudo random numbers
that are for all intents and purposes random. Here is an example
select (random(INT(1000 * (FLOAT(format(.#time,"hhmmss.sss")))))) from
sys_tables where count = 1
Marc wrote:
I just tested this 3 times and get the same numbers each time.
I am setting the Random number to 15 and created 10
The numbers I get are
0 0 5 0 5 3 8 2 10 14 each of the 3 times I tested this.
I hope I am missing something, I sent an update out several months
back that needs truley random numbers.
Marc
