R help - Aug 2010 - How to calc ratios base on current and previous row?

Hi all,

I want to calculate in each row a ratio based on number in the current row and
the previous row.
Is there a way to do this without for-loops because that is extremely slow.

	A	B	
[1]	2	2	
[2]	2	3 	
[3]	4	5,5
...

B2 = A2 + 0.5*B1


Thanks in advance.
Best
Marcus

Your question is unclear. If the calculation of B[3] now depends on the
newly calculated value of B[2], then the answer is no: you must loop. If
B[3] is based on the original B[2] value, then the answer is yes, and the
solution is just a matter of simple indexing, which I leave as an exercise.

-- Bert

On Fri, Aug 27, 2010 at 10:17 AM, Marcus Drescher <drescher@tum.de> wrote:
> Hi all,
>
> I want to calculate in each row a ratio based on number in the current row
> and the previous row.
> Is there a way to do this without for-loops because that is extremely slow.
>
>        A       B
> [1]     2       2
> [2]     2       3
> [3]     4       5,5
> ...
>
> B2 = A2 + 0.5*B1
>
>
> Thanks in advance.
> Best
> Marcus
>
> ______________________________________________
> R-help@r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
>
http://www.R-project.org/posting-guide.html<http://www.r-project.org/posting-guide.html>
> and provide commented, minimal, self-contained, reproducible code.
>
	[[alternative HTML version deleted]]

On 27-Aug-10 17:17:50, Marcus Drescher wrote:
> Hi all,
> I want to calculate in each row a ratio based on number in the current
> row and the previous row.
> Is there a way to do this without for-loops because that is extremely
> slow.
> 
>       A       B       
> [1]   2       2       
> [2]   2       3       
> [3]   4       5,5
> ...
> 
> B2 = A2 + 0.5*B1
> 
> Thanks in advance.
> Best
> Marcus
If n is the total number of rows, then :

  B[(2:n)] <- A[(2:n)] + 0.5*B[1:(n-1)]

should work! This assumes that A and B have been extracted as vectors,
or can be referred to separately as in a dataframe (say D), like

  D$B[(2:n)] <- D$A[(2:n)] + 0.5*D$B[1:(n-1)]

Ted.

--------------------------------------------------------------------
E-Mail: (Ted Harding) <Ted.Harding at manchester.ac.uk>
Fax-to-email: +44 (0)870 094 0861
Date: 27-Aug-10                                       Time: 18:34:27
------------------------------ XFMail ------------------------------

Hi Marcus,

I am guessing you are thinking in terms of Excel, where in column B,
you could enter the formula =(A2 + 0.5*B1) and just drag it down
however many cells you wanted.  In R, it would be expressed a bit
differently.  If k indexes your rows, I believe you want:

B[k] = A[k] + 0.5 * B[k - 1]

Unfortunately, if this is indeed what you mean, then as Bert said "If
the calculation of B[3] now depends on the newly calculated value of
B[2], then the answer is no: you must loop."

If calculating B is just some middle step to a larger goal, there may
be another way to your end goal that does not require a loop.  For
instance, suppose all you are really interested in is a single value
of B (perhaps the last one), you can use this formula where k = the
row number of the value of B you want:

A[k] + sum(A[1:(k - 1)] * (0.5 ^ ((k - 1):1)))

Best wishes,

Josh

On Fri, Aug 27, 2010 at 10:17 AM, Marcus Drescher <drescher at tum.de>
wrote:
> Hi all,
>
> I want to calculate in each row a ratio based on number in the current row
and the previous row.
> Is there a way to do this without for-loops because that is extremely slow.
>
> ? ? ? ?A ? ? ? B
> [1] ? ? 2 ? ? ? 2
> [2] ? ? 2 ? ? ? 3
> [3] ? ? 4 ? ? ? 5,5
> ...
>
> B2 = A2 + 0.5*B1
>
>
> Thanks in advance.
> Best
> Marcus
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>


-- 
Joshua Wiley
Ph.D. Student, Health Psychology
University of California, Los Angeles
http://www.joshuawiley.com/

For this particular case there is a nice shortcut (there is still looping , but
it is internal and quick):

a <- c(2,2,4, sample(1:10, 97, TRUE) )

b <- cumsum( (1/2)^(99:0)*a )/( (1/2)^(99:0) )

## compare

bb <- numeric(100)
bb[1] <- a[1]

for( i in 2:100 ) {
	bb[i] <- 1/2 * bb[i-1] + a[i]
}

all.equal( b, bb )

you will get some slight differences due to rounding error from the 2 methods.

this can be generalized to some degree, but if things get too complicated then
you may need to loop.

Hope this helps,

-- 
Gregory (Greg) L. Snow Ph.D.
Statistical Data Center
Intermountain Healthcare
greg.snow at imail.org
801.408.8111

> -----Original Message-----
> From: r-help-bounces at r-project.org [mailto:r-help-bounces at r-
> project.org] On Behalf Of Marcus Drescher
> Sent: Friday, August 27, 2010 11:18 AM
> To: r-help at r-project.org
> Subject: [R] How to calc ratios base on current and previous row?
> 
> Hi all,
> 
> I want to calculate in each row a ratio based on number in the current
> row and the previous row.
> Is there a way to do this without for-loops because that is extremely
> slow.
> 
> 	A	B
> [1]	2	2
> [2]	2	3
> [3]	4	5,5
> ...
> 
> B2 = A2 + 0.5*B1
> 
> 
> Thanks in advance.
> Best
> Marcus
> 
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-
> guide.html
> and provide commented, minimal, self-contained, reproducible code.