Re: [R] Avoiding for loops

[Boris Steipe]

If this code is slow it is not because you are using loops, but because you are 
dynamically building your vectors and lists and their size needs to change with 
each iteration causing significant unnecessary computational overhead. If you 
simply do something like

d <- numeric(m-1)
for (i in 1:m-1) {
 d[i] <- n - sum(x[1:i]) - i
}

for all of your loops, you will see already see very significant speedup. (If 
you look at my code formatting, and compare it with your own you may also 
benefit.) The bottom line: the point is not to avoid for-loops, but to speed up 
your code.

Nb. if you want to avoid loops for some aesthetic reason, read about apply() 
and its siblings,  and experiment with it. Of course, internally an apply() 
statement uses loops...

NNb: Do you know how to profile your code? How do you know which part of your 
code is actually slowing it down?



B.




On Oct 25, 2015, at 6:42 AM, Maram SAlem  wrote:

> Hi All,
> 
> I wonder if I can avoid the for() loop in any of the following loops.These
> loops are a part of a larger code which I'm trying to accelerate.
> 
> n=6
> m=4
> x<-c(0,1,1)
> 
> 1st loop
> 
> for (i in 1:m-1)
>  {
>  d[i]<- n- (sum(x[(1):(i)])) - i
>  }
> e<- n*(prod(d))
> 
> 
> 2nd loop
> 
> LD<-list()
>  for (i in 1:(m-1))
>  {
>  LD[[i]]<-seq(0,x[i],1)
>  }
> 
>  LD[[m]]<-seq(0,(n-m-sum(x)),1)
>  LED<-expand.grid (LD)
>  LED<-as.matrix(LED)
> 
> 
> 3rd loop
> 
> for (i in 1:(m-1))
> 
>  {
> 
>   h[i]<- choose(x[i],LED[j,i])
> 
>}
> 
> 
> 
> 4th loop
> 
> 
> for (i in 1:(m-1))
> 
> {
> 
>   lm[i]<-(sum(LED[j,1:i])) + i
> 
> }
> 
> 
> I appreciate if anyone has any suggestions or references.
> 
> 
> Thanks in advance.
> 
> 
> Maram Salem
> 
>   [[alternative HTML version deleted]]
> 
> __
> R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see
> 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.

__
R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see
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.

Re: [R] Avoiding for loops

[Berend Hasselman]

> On 25 Oct 2015, at 11:42, Maram SAlem  wrote:
> 
> Hi All,
> 
> I wonder if I can avoid the for() loop in any of the following loops.These
> loops are a part of a larger code which I'm trying to accelerate.
> 
> n=6
> m=4
> x<-c(0,1,1)
> 
> 1st loop
> 
> for (i in 1:m-1)
>   {
>   d[i]<- n- (sum(x[(1):(i)])) - i
>   }
>  e<- n*(prod(d))
> 
> 
On the basis of the other loops I presume you mean

for( in in 1:(m-1))

in stead of what you wrote.

You get the same result with 

d <- n - cumsum(x) - (1:(m-1))


Berend


>  2nd loop
> 
> LD<-list()
>   for (i in 1:(m-1))
>   {
>   LD[[i]]<-seq(0,x[i],1)
>   }
> 
>   LD[[m]]<-seq(0,(n-m-sum(x)),1)
>   LED<-expand.grid (LD)
>   LED<-as.matrix(LED)
> 
> 
> 3rd loop
> 
> for (i in 1:(m-1))
> 
>   {
> 
>h[i]<- choose(x[i],LED[j,i])
> 
> }
> 
> 
> 
> 4th loop
> 
> 
> for (i in 1:(m-1))
> 
>  {
> 
>lm[i]<-(sum(LED[j,1:i])) + i
> 
>  }
> 
> 
> I appreciate if anyone has any suggestions or references.
> 
> 
> Thanks in advance.
> 
> 
> Maram Salem
> 
>   [[alternative HTML version deleted]]
> 
> __
> R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see
> 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.

__
R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see
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.

[R] Avoiding for loops

[Maram SAlem]

Hi All,

I wonder if I can avoid the for() loop in any of the following loops.These
loops are a part of a larger code which I'm trying to accelerate.

n=6
m=4
x<-c(0,1,1)

1st loop

for (i in 1:m-1)
   {
   d[i]<- n- (sum(x[(1):(i)])) - i
   }
  e<- n*(prod(d))


  2nd loop

LD<-list()
   for (i in 1:(m-1))
   {
   LD[[i]]<-seq(0,x[i],1)
   }

   LD[[m]]<-seq(0,(n-m-sum(x)),1)
   LED<-expand.grid (LD)
   LED<-as.matrix(LED)


3rd loop

for (i in 1:(m-1))

   {

h[i]<- choose(x[i],LED[j,i])

 }



4th loop


 for (i in 1:(m-1))

  {

lm[i]<-(sum(LED[j,1:i])) + i

  }


I appreciate if anyone has any suggestions or references.


Thanks in advance.


Maram Salem

[[alternative HTML version deleted]]

__
R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see
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.

Re: [R] Avoiding for loops

[Maram SAlem]

Thanks a lot Boris and Berend.

I'll consider the brackets ((m-1) in every loop). In addition, I'll read
more on profiling my code. In fact,I'm using the apply () in another part
of my code.

Thanks again for helping.

Maram Salem

On 25 October 2015 at 14:26, Berend Hasselman  wrote:

>
> > On 25 Oct 2015, at 11:42, Maram SAlem  wrote:
> >
> > Hi All,
> >
> > I wonder if I can avoid the for() loop in any of the following
> loops.These
> > loops are a part of a larger code which I'm trying to accelerate.
> >
> > n=6
> > m=4
> > x<-c(0,1,1)
> >
> > 1st loop
> >
> > for (i in 1:m-1)
> >   {
> >   d[i]<- n- (sum(x[(1):(i)])) - i
> >   }
> >  e<- n*(prod(d))
> >
> >
> On the basis of the other loops I presume you mean
>
> for( in in 1:(m-1))
>
> in stead of what you wrote.
>
> You get the same result with
>
> d <- n - cumsum(x) - (1:(m-1))
>
>
> Berend
>
>
> >  2nd loop
> >
> > LD<-list()
> >   for (i in 1:(m-1))
> >   {
> >   LD[[i]]<-seq(0,x[i],1)
> >   }
> >
> >   LD[[m]]<-seq(0,(n-m-sum(x)),1)
> >   LED<-expand.grid (LD)
> >   LED<-as.matrix(LED)
> >
> >
> > 3rd loop
> >
> > for (i in 1:(m-1))
> >
> >   {
> >
> >h[i]<- choose(x[i],LED[j,i])
> >
> > }
> >
> >
> >
> > 4th loop
> >
> >
> > for (i in 1:(m-1))
> >
> >  {
> >
> >lm[i]<-(sum(LED[j,1:i])) + i
> >
> >  }
> >
> >
> > I appreciate if anyone has any suggestions or references.
> >
> >
> > Thanks in advance.
> >
> >
> > Maram Salem
> >
> >   [[alternative HTML version deleted]]
> >
> > __
> > R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see
> > 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.
>
>

[[alternative HTML version deleted]]

__
R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see
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.

Re: [R] Avoiding for loops

[Boris Steipe]

Sorry - I just noticed you actually have an error in your code: you had 
parentheses everywhere they were not needed and I overlooked you had not put 
them where they actually are needed. It has to be for (i in 1:(m-1)) ..., not 
as you wrote. I'm sure you'll understand the difference.

d <- numeric(m-1)
for (i in 1:(m-1)) {
d[i] <- n - sum(x[1:i]) - i
}

B.

On Oct 25, 2015, at 8:13 AM, Boris Steipe  wrote:

> If this code is slow it is not because you are using loops, but because you 
> are dynamically building your vectors and lists and their size needs to 
> change with each iteration causing significant unnecessary computational 
> overhead. If you simply do something like
> 
> d <- numeric(m-1)
> for (i in 1:m-1) {
>  d[i] <- n - sum(x[1:i]) - i
> }
> 
> for all of your loops, you will see already see very significant speedup. (If 
> you look at my code formatting, and compare it with your own you may also 
> benefit.) The bottom line: the point is not to avoid for-loops, but to speed 
> up your code.
> 
> Nb. if you want to avoid loops for some aesthetic reason, read about apply() 
> and its siblings,  and experiment with it. Of course, internally an apply() 
> statement uses loops...
> 
> NNb: Do you know how to profile your code? How do you know which part of your 
> code is actually slowing it down?
> 
> 
> 
> B.
> 
> 
> 
> 
> On Oct 25, 2015, at 6:42 AM, Maram SAlem  wrote:
> 
>> Hi All,
>> 
>> I wonder if I can avoid the for() loop in any of the following loops.These
>> loops are a part of a larger code which I'm trying to accelerate.
>> 
>> n=6
>> m=4
>> x<-c(0,1,1)
>> 
>> 1st loop
>> 
>> for (i in 1:m-1)
>>  {
>>  d[i]<- n- (sum(x[(1):(i)])) - i
>>  }
>> e<- n*(prod(d))
>> 
>> 
>> 2nd loop
>> 
>> LD<-list()
>>  for (i in 1:(m-1))
>>  {
>>  LD[[i]]<-seq(0,x[i],1)
>>  }
>> 
>>  LD[[m]]<-seq(0,(n-m-sum(x)),1)
>>  LED<-expand.grid (LD)
>>  LED<-as.matrix(LED)
>> 
>> 
>> 3rd loop
>> 
>> for (i in 1:(m-1))
>> 
>>  {
>> 
>>   h[i]<- choose(x[i],LED[j,i])
>> 
>>}
>> 
>> 
>> 
>> 4th loop
>> 
>> 
>> for (i in 1:(m-1))
>> 
>> {
>> 
>>   lm[i]<-(sum(LED[j,1:i])) + i
>> 
>> }
>> 
>> 
>> I appreciate if anyone has any suggestions or references.
>> 
>> 
>> Thanks in advance.
>> 
>> 
>> Maram Salem
>> 
>>  [[alternative HTML version deleted]]
>> 
>> __
>> R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see
>> 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.
> 

__
R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see
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.

Re: [R] avoiding for loops

[R. Michael Weylandt]

Usually you can just use cor() and it will do all the possibilities directly:

x - matrix(rnorm(100), ncol = 10)
cor(x)

But that works on the columns, so you'll need to transpose things if
you want all possible row combinations: cor(t(x))

Hope this helps,
Michael

On Fri, Apr 6, 2012 at 9:57 AM, Cserháti Mátyás cs_ma...@yahoo.com wrote:
 Hello everyone,

 My name is Matthew and I'm new to this list. greetings to everyone.
 Sorry if I'm asking an old question, but I have an m x n matrix where the 
 rows are value profiles and the columns are conditions.
 What I want to do is calculate the correlation between all possible pairs of 
 rows.

 That is, if there are 10 rows in my matrix, then I want to calculate 10 x  10 
 = 100 correlation values (all against all).
 Now R is slow when I use two for loops.
 What kind of other function or tool can I use to get the job done more 
 speedily?
 I've heard of tapply, lapply, etc. and by and aggregate.

 Any kind of help is gladly appreciated.

 Thanks,


 Matthew


[Deleted the unnecessary digest]

__
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
and provide commented, minimal, self-contained, reproducible code.

[R] avoiding for loops

[Ed Siefker]

I have data that looks like this:

 df1
  group id
1   red  A
2   red  B
3   red  C
4  blue  D
5  blue  E
6  blue  F


I want a list of the groups containing vectors with the ids.I am
avoiding subset(), as it is
only recommended for interactive use.  Here's what I have so far:

df1 - data.frame(group=c(red, red, red, blue, blue,
blue), id=c(A, B, C, D, E, F))

groups - levels(df1$group)
byid - lapply(groups, ==, df1$group)
groupIDX - lapply(byid, which)

 groupIDX
[[1]]
[1] 4 5 6

[[2]]
[1] 1 2 3



This gives me a list of the indices for each group.  I want to subset
df1 based on this list.
If I want just one group I can do this:

 df1[groupIDX[[1]],]$id
[1] D E F


What sort of statement should I use if I want a result like:
[[1]]
[1] D E F
Levels: A B C D E F

[[2]]
[1] A B C
Levels: A B C D E F


So far, I've used a for loop.  Can I express this with apply statements?

groupIDs - list(1:length(groupIDX))
groupData-
for (i in 1:length(groupIDX)) {
groupIDs[[i]] - df1[groupIDX[[i]],]$id
}

__
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
and provide commented, minimal, self-contained, reproducible code.

Re: [R] avoiding for loops

[Richard M. Heiberger]

 df1 - data.frame(group=c(red, red, red, blue, blue,
blue), id=c(A, B, C, D, E, F))
df1 - data.frame(group=c(red, red, red, blue, blue,
+ blue), id=c(A, B, C, D, E, F))
 df1
  group id
1   red  A
2   red  B
3   red  C
4  blue  D
5  blue  E
6  blue  F
 with(df1, split(id, group))
$blue
[1] D E F
Levels: A B C D E F

$red
[1] A B C
Levels: A B C D E F




On Sun, Mar 25, 2012 at 4:46 PM, Ed Siefker ebs15...@gmail.com wrote:

 I have data that looks like this:

  df1
  group id
 1   red  A
 2   red  B
 3   red  C
 4  blue  D
 5  blue  E
 6  blue  F


 I want a list of the groups containing vectors with the ids.I am
 avoiding subset(), as it is
 only recommended for interactive use.  Here's what I have so far:

 df1 - data.frame(group=c(red, red, red, blue, blue,
 blue), id=c(A, B, C, D, E, F))

 groups - levels(df1$group)
 byid - lapply(groups, ==, df1$group)
 groupIDX - lapply(byid, which)

  groupIDX
 [[1]]
 [1] 4 5 6

 [[2]]
 [1] 1 2 3



 This gives me a list of the indices for each group.  I want to subset
 df1 based on this list.
 If I want just one group I can do this:

  df1[groupIDX[[1]],]$id
 [1] D E F


 What sort of statement should I use if I want a result like:
 [[1]]
 [1] D E F
 Levels: A B C D E F

 [[2]]
 [1] A B C
 Levels: A B C D E F


 So far, I've used a for loop.  Can I express this with apply statements?

 groupIDs - list(1:length(groupIDX))
 groupData-
 for (i in 1:length(groupIDX)) {
groupIDs[[i]] - df1[groupIDX[[i]],]$id
}

 __
 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.htmlhttp://www.r-project.org/posting-guide.html
 and provide commented, minimal, self-contained, reproducible code.


[[alternative HTML version deleted]]

__
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
and provide commented, minimal, self-contained, reproducible code.

[R] Avoiding two loops

[sudhir cr]

Hello,

I have a R code for doing convolution of two functions:

convolveSlow - function(x, y) {
nx - length(x); ny - length(y)
xy - numeric(nx + ny - 1)
for(i in seq(length = nx)) {
xi - x[[i]]
for(j in seq(length = ny)) {
ij - i+j-1
xy[[ij]] - xy[[ij]] + xi * y[[j]]
}
}
xy
}

How do I reduce the 2 loops so that I can run the  code faster?

Thank you
San

[[alternative HTML version deleted]]

__
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
and provide commented, minimal, self-contained, reproducible code.

Re: [R] Avoiding two loops

[Dirk Eddelbuettel]

On 4 February 2011 at 14:03, sudhir cr wrote:
| Hello,
| 
| I have a R code for doing convolution of two functions:
| 
| convolveSlow - function(x, y) {
| nx - length(x); ny - length(y)
| xy - numeric(nx + ny - 1)
| for(i in seq(length = nx)) {
| xi - x[[i]]
| for(j in seq(length = ny)) {
| ij - i+j-1
| xy[[ij]] - xy[[ij]] + xi * y[[j]]
| }
| }
| xy
| }
| 
| How do I reduce the 2 loops so that I can run the  code faster?

Maybe by reading the answer to the _exact same question_ you appear to have
asked on SO yesterday?

http://stackoverflow.com/questions/4894506/avoid-two-for-loops-in-r

Dirk

| 
| Thank you
| San
| 
|   [[alternative HTML version deleted]]
| 
| __
| 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
| and provide commented, minimal, self-contained, reproducible code.

-- 
Dirk Eddelbuettel | e...@debian.org | http://dirk.eddelbuettel.com

__
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
and provide commented, minimal, self-contained, reproducible code.

Re: [R] Avoiding two loops

[Petr Savicky]

On Fri, Feb 04, 2011 at 02:03:22PM -0500, sudhir cr wrote:
 Hello,
 
 I have a R code for doing convolution of two functions:
 
 convolveSlow - function(x, y) {
 nx - length(x); ny - length(y)
 xy - numeric(nx + ny - 1)
 for(i in seq(length = nx)) {
 xi - x[[i]]
 for(j in seq(length = ny)) {
 ij - i+j-1
 xy[[ij]] - xy[[ij]] + xi * y[[j]]
 }
 }
 xy
 }
 
 How do I reduce the 2 loops so that I can run the  code faster?

Hello:

Convolution of two vectors may be computed also using matrix reshaping
without a loop. For example 

  convolution - function(x, y)
  {
  # more efficient if length(x) = length(y)
  m - length(x)
  n - length(y)
  zero - matrix(0, nrow=n, ncol=n)
  a - rbind(x %o% y, zero)
  k - m + n - 1
  b - matrix(c(a)[1:(n*k)], nrow=k, ncol=n)
  rowSums(b)
  }

Testing this on computing the product of the polynomials (1+t)^4 (1+t)^3
yields

  x - choose(4, 0:4)
  y - choose(3, 0:3)
  convolution(x, y)

  [1]  1  7 21 35 35 21  7  1

which is the same as choose(7, 0:7).

See also ?convolve.

Hope this helps.

Petr Savicky.

__
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
and provide commented, minimal, self-contained, reproducible code.

[R] Avoiding for loops

[Noah Silverman]

Hi,

I'm trying to normalize some data.
My data is organized by groups.  I want to normalize PER GROUP as 
opposed to over the entire data set.


The current double loop that I'm using takes almost an hour to run on 
about 30,000 rows of data in 2,500 groups.


I'm currently doing this:

-
for(group in unique(data$group)){
sum_V1 - sum(data$V1[data$group == group])

for(subject in data$subject[data$group == group]){
data$V1_norm[(data$group == group  data$subject == subject)] 
- data$V1[(data$group == group  data$subject == subject)] / sum_V1

}
}
-

Can anyone point me to a faster way to do this in R.

Thanks!

-N

__
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
and provide commented, minimal, self-contained, reproducible code.

Re: [R] Avoiding for loops

[Dimitris Rizopoulos]

you could try something along these lines:

data - data.frame(y = rnorm(100), group = rep(1:10, each = 10))

data$sum - ave(data$y, data$group, FUN = sum)
data$norm.y - data$y / data$sum
data


I hope it helps.

Best,
Dimitris


Noah Silverman wrote:

Hi,

I'm trying to normalize some data.
My data is organized by groups.  I want to normalize PER GROUP as 
opposed to over the entire data set.


The current double loop that I'm using takes almost an hour to run on 
about 30,000 rows of data in 2,500 groups.


I'm currently doing this:

-
for(group in unique(data$group)){
sum_V1 - sum(data$V1[data$group == group])

for(subject in data$subject[data$group == group]){
data$V1_norm[(data$group == group  data$subject == subject)] - 
data$V1[(data$group == group  data$subject == subject)] / sum_V1

}
}
-

Can anyone point me to a faster way to do this in R.

Thanks!

-N

__
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

and provide commented, minimal, self-contained, reproducible code.



--
Dimitris Rizopoulos
Assistant Professor
Department of Biostatistics
Erasmus University Medical Center

Address: PO Box 2040, 3000 CA Rotterdam, the Netherlands
Tel: +31/(0)10/7043478
Fax: +31/(0)10/7043014

__
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
and provide commented, minimal, self-contained, reproducible code.

Re: [R] Avoiding for loops

[Peter Dalgaard]

Dimitris Rizopoulos wrote:
 you could try something along these lines:
 
 data - data.frame(y = rnorm(100), group = rep(1:10, each = 10))
 
 data$sum - ave(data$y, data$group, FUN = sum)
 data$norm.y - data$y / data$sum
 data

.. or even

transform(data, norm=ave(y, group, FUN = function(x) x/sum(x)))

 
 I hope it helps.
 
 Best,
 Dimitris
 
 
 Noah Silverman wrote:
 Hi,

 I'm trying to normalize some data.
 My data is organized by groups.  I want to normalize PER GROUP as
 opposed to over the entire data set.

 The current double loop that I'm using takes almost an hour to run on
 about 30,000 rows of data in 2,500 groups.

 I'm currently doing this:

 -
 for(group in unique(data$group)){
 sum_V1 - sum(data$V1[data$group == group])

 for(subject in data$subject[data$group == group]){
 data$V1_norm[(data$group == group  data$subject == subject)]
 - data$V1[(data$group == group  data$subject == subject)] / sum_V1
 }
 }
 -

 Can anyone point me to a faster way to do this in R.

 Thanks!

 -N

 __
 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
 and provide commented, minimal, self-contained, reproducible code.

 


-- 
   O__   Peter Dalgaard Øster Farimagsgade 5, Entr.B
  c/ /'_ --- Dept. of Biostatistics PO Box 2099, 1014 Cph. K
 (*) \(*) -- University of Copenhagen   Denmark  Ph:  (+45) 35327918
~~ - (p.dalga...@biostat.ku.dk)  FAX: (+45) 35327907

__
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
and provide commented, minimal, self-contained, reproducible code.

Re: [R] Avoiding for loops

[Noah Silverman]

Wow,

That's nice.

Should work well, but I just realized that I missed something in 
explaining my code.

I need to calculate the exp function on X

so it should be

exp(x) / sum(exp(x)) for each group

I tried this with:

foo - ave(rawdata$foo,rawdata$code,FUN=function(x) exp(x) / sum(exp(x)))

That didn't work.  I got a lot of NaN

-N

On 11/2/09 3:30 AM, Peter Dalgaard wrote:
 Dimitris Rizopoulos wrote:

 you could try something along these lines:

 data- data.frame(y = rnorm(100), group = rep(1:10, each = 10))

 data$sum- ave(data$y, data$group, FUN = sum)
 data$norm.y- data$y / data$sum
 data
  
 .. or even

 transform(data, norm=ave(y, group, FUN = function(x) x/sum(x)))


 I hope it helps.

 Best,
 Dimitris


 Noah Silverman wrote:
  
 Hi,

 I'm trying to normalize some data.
 My data is organized by groups.  I want to normalize PER GROUP as
 opposed to over the entire data set.

 The current double loop that I'm using takes almost an hour to run on
 about 30,000 rows of data in 2,500 groups.

 I'm currently doing this:

 -
 for(group in unique(data$group)){
  sum_V1- sum(data$V1[data$group == group])

  for(subject in data$subject[data$group == group]){
  data$V1_norm[(data$group == group  data$subject == subject)]
 - data$V1[(data$group == group  data$subject == subject)] / sum_V1
  }
 }
 -

 Can anyone point me to a faster way to do this in R.

 Thanks!

 -N

 __
 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
 and provide commented, minimal, self-contained, reproducible code.


  



[[alternative HTML version deleted]]

__
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
and provide commented, minimal, self-contained, reproducible code.

[R] Avoiding FOR loops

[dxc13]

useR's,

I would like to know if there is a way to avoid using FOR loops to perform
the below calculation. 

Consider the following data:

 x
 [,1] [,2] [,3]
[1,]4   111
[2,]192
[3,]733
[4,]364
[5,]685

 xk
 Var1  Var2 Var3
1   -0.25  1.75  0.5
20.75  1.75  0.5
31.75  1.75  0.5
42.75  1.75  0.5
53.75  1.75  0.5
64.75  1.75  0.5
75.75  1.75  0.5
86.75  1.75  0.5
97.75  1.75  0.5
10  -0.25  2.75  0.5

Here, X is a matrix of 3 variables in which each is of size 5 and XK are
some values that correspond to each variable.  For each variable, I want to
do:

|Xi - xkj|   where i = 1 to 3 and j = 1 to 10

It looks as if a double FOR loop would work, but can the apply function
work?  Or some other function that is shorter than a FOR loop?  Thank you, I
hope this makes sense.

Derek


-- 
View this message in context: 
http://www.nabble.com/Avoiding-FOR-loops-tp14656517p14656517.html
Sent from the R help mailing list archive at Nabble.com.

__
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
and provide commented, minimal, self-contained, reproducible code.

Re: [R] Avoiding FOR loops

[Charilaos Skiadas]

On Jan 6, 2008, at 7:55 PM, dxc13 wrote:


 useR's,

 I would like to know if there is a way to avoid using FOR loops to  
 perform
 the below calculation.

 Consider the following data:

snip
 Here, X is a matrix of 3 variables in which each is of size 5 and  
 XK are
 some values that correspond to each variable.  For each variable, I  
 want to
 do:

 |Xi - xkj|   where i = 1 to 3 and j = 1 to 10

That should be i=1 to 5 I take it?

If I understand what you want to do, then the outer function is the key:

lapply(1:3, function(i) { outer(x[,i], xk[,i],  -) } )

This should land you with a list of three 5x10 tables

 It looks as if a double FOR loop would work, but can the apply  
 function
 work?  Or some other function that is shorter than a FOR loop?   
 Thank you, I
 hope this makes sense.

 Derek


Haris Skiadas
Department of Mathematics and Computer Science
Hanover College

__
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
and provide commented, minimal, self-contained, reproducible code.