R help - Nov 2006 - one row matrix

Hi,

when I assign a one row matrix to another variable, then somehow R
automatically convert that varaible into  a list.

For example:
> a = matrix (12, 3, 4)
> b = a[1,]
> b
[1] 12 12 12 12

Is there a way to enable R automatically make b as a one row matrix , rather
than explicitly assign like matrix (b, ncol=1).   I have come across a
couple of time, that when I try to check how many rows of the computed
result matrix - nrow(,,,), because there is only one row, the program failed
as R tried to interpret it as a list.

Can some one help?

-- 
Waverley @ Palo Alto

	[[alternative HTML version deleted]]

On Fri, 2006-11-03 at 14:12 -0800, Waverley wrote:
> Hi,
> 
> when I assign a one row matrix to another variable, then somehow R
> automatically convert that varaible into  a list.
A 'vector', not a 'list'
> For example:
> 
> > a = matrix (12, 3, 4)
> > b = a[1,]
> > b
> [1] 12 12 12 12
> 
> Is there a way to enable R automatically make b as a one row matrix ,
rather
> than explicitly assign like matrix (b, ncol=1).   I have come across a
> couple of time, that when I try to check how many rows of the computed
> result matrix - nrow(,,,), because there is only one row, the program
failed
> as R tried to interpret it as a list.
> 
> Can some one help?
This is actually a FAQ:

http://cran.r-project.org/doc/FAQ/R-FAQ.html#Why-do-my-matrices-lose-dimensions_003f

HTH,

Marc Schwartz

On 11/3/06, Waverley <waverley.paloalto at gmail.com>
wrote:
> Hi,
> For example:
>
> > a = matrix (12, 3, 4)
> > b = a[1,]
> > b
> [1] 12 12 12 12
>
It doesn't have anything to do with one-row matrices, at least in your
example,
since that isn't what you made. Instead, you created a matrix with 3 rows,
4 columns, and all values equal to 12.
> a <- matrix(12, 3, 4)
> a
     [,1] [,2] [,3] [,4]
[1,]   12   12   12   12
[2,]   12   12   12   12
[3,]   12   12   12   12
> a[1,]
[1] 12 12 12 12

You probably intended
> a <- matrix(c(12, 3, 4), nrow=1)
> a
     [,1] [,2] [,3]
[1,]   12    3    4

But to solve what I think you were actually asking, even though it
wasn't reflected in your example... By default, R drops unused
dimensions when subsetting. You can override this behavior
> b <- a[1, ,drop=FALSE]
> b
     [,1] [,2] [,3]
[1,]   12    3    4

See:
?subset
help('[')
?matrix

Sarah
-- 
Sarah Goslee
http://www.functionaldiversity.org