[R] reshape

[Thomas Lumley]

On Mon, 9 Sep 2002, Ngayee J Law wrote:

> Hi,
>
> I have a problem using the function 'reshape' for the following dataset:
>
> input:
> X    Y    Z    T.U.1     T.U.2     T.V.1      T.V.2
> 1.00 1.00 A    24.94     24.74     34.65      35.33
> 1.00 3.00 A    26.03     25.87     25.76      26.12
> 1.00 1.00 B    24.76     24.63     34.34      34.48
> 1.00 3.00 B    26.15     25.98     25.96      25.66
> 2.00 1.00 A    24.40     24.26     34.55      36.53
> 2.00 4.00 A    24.55     24.40     35.52      39.61
> 2.00 1.00 B    25.00     24.76     35.32      36.67
> 2.00 4.00 B    25.96     25.70     26.00      25.62
>
> target output:
> X    Y    Z    T    Time W
> 1.00 1.00 A    U    1    24.94
> 1.00 1.00 A    U    2    24.74
> 1.00 1.00 A    V    1    34.65
> 1.00 1.00 A    V    2    35.33
> 1.00 3.00 A    U    1    26.03
> 1.00 3.00 A    U    2    25.87
> 1.00 3.00 A    V    1    25.76
> 1.00 3.00 A    V    2    26.12
> 1.00 1.00 B    U    1    24.76
> 1.00 1.00 B    U    2    24.63
> 1.00 1.00 B    V    1    34.34
> 1.00 1.00 B    V    2    34.48
> 1.00 3.00 B    U    1    26.15
> 1.00 3.00 B    U    2    25.98
> 1.00 3.00 B    V    1    25.96
> 1.00 3.00 B    V    2    25.66
> 2.00 1.00 A    U    1    24.40
> 2.00 1.00 A    U    2    24.26
> 2.00 1.00 A    V    1    34.55
> 2.00 1.00 A    V    2    36.53
> 2.00 4.00 A    U    1    24.55
> 2.00 4.00 A    U    2    24.40
> 2.00 4.00 A    V    1    35.52
> 2.00 4.00 A    V    2    39.61
> 2.00 1.00 B    U    1    25.00
> 2.00 1.00 B    U    2    24.76
> 2.00 1.00 B    V    1    35.32
> 2.00 1.00 B    V    2    36.67
> 2.00 4.00 B    U    1    25.96
> 2.00 4.00 B    U    2    25.70
> 2.00 4.00 B    V    1    26.00
> 2.00 4.00 B    V    2    25.62
>
>
> There are more than 1 varying for the idvar (X, Y, Z) and there are two
> variables
> for the varying option (U, V). Any ideas how to use reshape for this
> problem?

The simplest way is to do it in two steps
Suppose the data are in data frame `a'
    b<-reshape(a,direction="long",
		varying=list(c("T.U.1","T.U.2"), c("T.V.1","T.V.2")),
		v.names=c("U","V"))
looks like
    X Y Z time     U     V id
1.1 1 1 A    1 24.94 34.65  1
2.1 1 3 A    1 26.03 25.76  2
3.1 1 1 B    1 24.76 34.34  3
4.1 1 3 B    1 26.15 25.96  4
5.1 2 1 A    1 24.40 34.55  5
6.1 2 4 A    1 24.55 35.52  6
7.1 2 1 B    1 25.00 35.32  7
8.1 2 4 B    1 25.96 26.00  8
1.2 1 1 A    2 24.74 35.33  1
2.2 1 3 A    2 25.87 26.12  2
3.2 1 1 B    2 24.63 34.48  3
4.2 1 3 B    2 25.98 25.66  4
5.2 2 1 A    2 24.26 36.53  5
6.2 2 4 A    2 24.40 39.61  6
7.2 2 1 B    2 24.76 36.67  7
8.2 2 4 B    2 25.70 25.62  8

Now, you want to reshape again, but specify that the time variable is
called "T", with values "U" and "V". You also want to drop the id
variable.


reshape(b[,-7],direction="long",
        varying=list(c("U","V")), v.names="W",
 	times=c("U","V"), timevar="T")

Again, you probably want to drop the `id' variable and if you care
about the order of rows, sort on X and Y.

	-thomas

Thomas Lumley			Asst. Professor, Biostatistics
tlumley at u.washington.edu	University of Washington, Seattle


-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._