After slogging through lots of posts about a poorly defined request, I am
left wondering if I missed the original sender properly explaining what
THEY mean by a truth table and what it should look like.
Some here seem to know (or are guessing) that the request is to make all
combinations of TRUE and FALSE for N columns in a data.frame and some for
an indefinite value of N. Some others may also want to throw in additional
columns that reflect a logical AND operation and perhaps others.
So I calmly request someone tell us what the real request is so I can
evaluate if anything said here makes much sense in answering the real
request.
As I see it, if you have 2 columns, there are four possible combination in
what amounts to a 4x2 matrix. If your mailer allows my text to be seen as
intended, the following shows combinations starting with F, albeit a table
starting with T is equivalent in terms of meaning:
FFFTTFTT
For an N=3 column it gets more rows using binary notation with T=0 and F=1
so 8 rows.
000001010011100101110111
The trend becomes clear that the number of rows is 2**N power so a simple
approach (albeit there are other ways shown that may be simpler to code
using existing software) is to note the pattern. The first column requires
2**N items alternating every (2**N)/2 times. Meaning if N=5 then you want
32 rows in the result with 16 units of F and then 16 units of T, or vice
versa. The R function that does this easily (as part of a loop perhaps) is
rep() and sample code (hopefully blank lines keep it from getting wrapped
funny is something like this that can be simplified:
N <- 5
rows <- 2**N
TF <- data.frame(index=1:rows)
for (ind in rev(2**(N:1))) { TF <- cbind(TF, rep(c(TRUE, FALSE),
each=rows/ind, length.out=rows)) }
names(TF) <- c("index", paste("col", 1:N, sep=""))
The above uses rep() repeatedly to produce runs of TRUE and FALSE of
decreasing size and keeps concatenating them to an existing data.frame with
cbind(). The result is a column with 16 TRUE followed by 16 FALSE then
another column with 8 by 8 and repeated again as 8 by 8. The next column
alternates in groups of 4 then the next in groups of two and finally
alternating in "groups" of 1.
Obviously this can be wrapped up in a function that takes N as an argument
and makes an arbitrary N column construct with 2**N rows as described and
this may be what is wanted for the main table. I threw this together
rapidly and I am sure can improve it so column names are created as
appropriate.
For example, rather than cbind, the following would work well too:
colnm <- ...TF[colnm] <- rep(c(TRUE, FALSE), each=rows/ind,
length.out=rows)
But the question is whether this makes what is wanted, or needs something
more like columns that represent whether the OR or AND or some other
boolean function of N boolean items is TRUE or FALSE.
I repeat, the above analysis does not suggest other supplied solutions are
bad or wrong, just a suggestion of how fairly simple functionality in R can
do what is wanted. Of course, if something else is wanted, we are all
wasting our time answering. I waited a while hoping not to need to or to
reply to an actual question I know how to deal with.
-----Original Message-----
From: Jeff Newmiller <jdnewmil at dcn.davis.ca.us>
To: Bert Gunter <bgunter.4567 at gmail.com>; Ebert, Timothy Aaron <
tebert at ufl.edu>
Cc: r-help at r-project.org <r-help at r-project.org>; Paul Bernal <
paulbernal07 at gmail.com>
Sent: Sun, Mar 13, 2022 5:17 am
Subject: Re: [R] Is there a Truth Table Generator in R?
There are 2^(2^length(tt)) possible "truth" vectors for the inputs defined
in tt. AND-ing all of the inputs only gives one of those possibilities.
Some popular named cases for 2 inputs are shown here [1], but it is common
to use combinations of !, & and | to specify a particular truth vector.
There is also the problem of reverse-engineering such a boolean expeession
[2] in simplest form from a given truth vector, but I don't know if anyone
has implemented such algorithms in R.
[1] https://en.wikipedia.org/wiki/Truth_table
[2] https://en.wikipedia.org/wiki/Karnaugh_maps
On March 12, 2022 2:17:32 PM PST, Bert Gunter <bgunter.4567 at gmail.com>
wrote:
...
tt$truth <- tt$A & tt$B & tt$C
to evaluate the outcome of expand.grid.
or, as I said,
tt$truth <- apply(tt,1, all)
which works for any number of columns in tt.
Bert Gunter
"The trouble with having an open mind is that people keep coming along
and sticking things into it."
-- Opus (aka Berkeley Breathed in his "Bloom County" comic strip )
On Sat, Mar 12, 2022 at 2:02 PM Ebert,Timothy Aaron <tebert at ufl.edu>
To the end of Jeff's program add
tt$truth <- tt$A & tt$B & tt$C
to evaluate the outcome of expand.grid.
Tim
-----Original Message-----
From: R-help <r-help-bounces at r-project.org> On Behalf Of Jeff Newmiller
Sent: Saturday, March 12, 2022 12:05 PM
To: r-help at r-project.org; Paul Bernal <paulbernal07 at gmail.com>; R <
Subject: Re: [R] Is there a Truth Table Generator in R?
[External Email]
both <- c( FALSE, TRUE )
tt <- expand.grid( C = both
, B = both
, A = both
)
tt <- tt[, 3:1 ]
On March 12, 2022 8:42:28 AM PST, Paul Bernal <paulbernal07 at gmail.com>
Dear friends,
Hope you are doing great. I have been searching for a truth table
generator in R, but everything I find has a Python implementation
Maybe there is in fact a truth table generator in R, but I am not
searching in the right places?
Any help and/or guidance will be greatly appreciated.
Best regards,
Paul
[[alternative HTML version deleted]]