R help - Jun 2005 - merge large matrices

Dear List

I have two large matrices A and B. Both have the same dimensions, let's 
say 20k x 30k. About half the cells of B are missing. Now I'm looking 
for an efficient way to merge them, so that the missing values in B are 
replaced by the corresponding values of A.

Matrix A
	[,1]	[,2]	[,3]
[1,]	1	2	3
[2,]	4	5	6

merged with Matrix B

	[,1]	[,2]	[,3]
[1,]	10	NA	NA
[2,]	NA	50	60

equals

	[,1]	[,2]	[,3]
[1,]	10	2	3
[2,]	4	50	60

One way to do this, is  of course looping through all the cells, 
checking for NAs and then replacing them with the corresponding values. 
But this is way too slow for my application. There must be a more 
efficient way.
Does R provide any functions for this?

Cheers
Stefan

On 6/2/05, Stefan Mischke <mischke at sozpsy.unizh.ch>
wrote:
> Dear List
> 
> I have two large matrices A and B. Both have the same dimensions, let's
> say 20k x 30k. About half the cells of B are missing. Now I'm looking
> for an efficient way to merge them, so that the missing values in B are
> replaced by the corresponding values of A.
How about:
B[is.na(B)] <- A[is.na(B)]


Sarah

-- 
Sarah Goslee
http://www.stringpage.com

On 6/2/2005 4:19 PM, Stefan Mischke wrote:
> Dear List
> 
> I have two large matrices A and B. Both have the same dimensions, let's
> say 20k x 30k. About half the cells of B are missing. Now I'm looking 
> for an efficient way to merge them, so that the missing values in B are 
> replaced by the corresponding values of A.
> 
> Matrix A
> 	[,1]	[,2]	[,3]
> [1,]	1	2	3
> [2,]	4	5	6
> 
> merged with Matrix B
> 
> 	[,1]	[,2]	[,3]
> [1,]	10	NA	NA
> [2,]	NA	50	60
> 
> equals
> 
> 	[,1]	[,2]	[,3]
> [1,]	10	2	3
> [2,]	4	50	60
> 
> One way to do this, is  of course looping through all the cells, 
> checking for NAs and then replacing them with the corresponding values. 
> But this is way too slow for my application. There must be a more 
> efficient way.
> Does R provide any functions for this?
As long as the dimensions are the same:

replace <- is.na(B)
B[replace] <- A[replace]

should work.

Duncan Murdoch

Thanks a lot!
Both versions seem to work equally efficient.
N8,
Stefan


> replace <- is.na(B)
> B[replace] <- A[replace]
>
> Duncan Murdoch
> B[is.na(B)] <- A[is.na(B)]
>
> Sarah