reshape with two time variables
On Fri, Jan 30, 2009 at 5:57 PM, Neil Stewart
<neil.stewart at warwick.ac.uk> wrote:
I have a data frame in wide format that I'd like to convert to long format.
For example, in wide format I have:
id A1B1 A1B2 A2B1 A2B2
1 1 400 475 420 510
2 2 390 500 470 472
3 3 428 512 555 610
4 4 703 787 801 822
5 5 611 634 721 705
6 6 543 522 612 788
7 7 411 488 506 623
8 8 654 644 711 795
A is one repeated-measures variable with levels 1 and 2. B is a second
repeated-measures variable also with levels 1 and 2. I'd like to end up
with:
id A B value
1 1 1 1 400
2 2 1 1 390
3 3 1 1 428
4 4 1 1 703
5 5 1 1 611
6 6 1 1 543
7 7 1 1 411
8 8 1 1 654
9 1 1 2 475
10 2 1 2 500
11 3 1 2 512
12 4 1 2 787
13 5 1 2 634
14 6 1 2 522
15 7 1 2 488
16 8 1 2 644
17 1 2 1 420
18 2 2 1 470
19 3 2 1 555
20 4 2 1 801
21 5 2 1 721
22 6 2 1 612
23 7 2 1 506
24 8 2 1 711
25 1 2 2 510
26 2 2 2 472
27 3 2 2 610
28 4 2 2 822
29 5 2 2 705
30 6 2 2 788
31 7 2 2 623
32 8 2 2 795
I've been working with reshape and with the reshape package, but with no
success for the two-time-variable case. Can anyone help please?
If you're using melt from the reshape package, once you've melted the data, you'll need to do something like: m$a <- as.numeric(substr(m$variable, 1, 1)) m$b <- as.numeric(substr(m$variable, 3, 3)) m$variable <- NULL Hadley