Skip to content

Reformat meteorological data

2 messages · Toby Gass, Hadley Wickham

Toby Gass
# 
Dear HelpeRs:

I have a data set in the following format,
which will be familiar to those of you working
with NCDC climate data.

Example:

Year <- rep(1:3, each = 3)
Year <- c(Year,Year)
ID <- rep(1:2, each = 9)
Jan <- runif(18, min = 0, max = 20)
Feb <- runif(18, min = 0, max = 20)
Mar <- runif(18, min = 0, max = 20)
var <- gl(3,1,18,label = c("snow","rain","temp"))
data <-data.frame(row.names = c(1:18),ID, Year, var, Jan, 
Feb, Mar)


The actual dataset has 12 months, about 100 years, 6 ID's,
many more levels in "var", and includes NA's.


I would like to convert the data to the following format:

ID Year Month Rain Snow Temp
1     1     Jan     value  value value
1     1     Feb     value value value
1     1     Mar     value value value
. . . . .
. . . . .
. . . . .
1     2     Jan     value value value
1     2     Feb     value value value
1     2     Mar     value value value
. . . . . .
. . . . . .
. . . . . .
1     3     Jan     value value value
1     3     Feb     value value value
1     3     Mar     value value value
. . . . . .
. . . . . .
2     1     Jan     value value value
2     1     Feb     value value value
2     1     Mar     value value value
. . . . .
. . . . .
. . . . .
2     2     Jan     value value value
2     2     Feb     value value value
2     2     Mar     value value value

and so on.

I'd appreciate some code that would help accomplish this.
Since I'm not an R expert, code that is somewhat transparent
might be more helpful to me than the shortest possible 
option.

Thank you.


Toby Gass
Department of Forest, Rangeland, and Watershed Stewardship
Warner College of Natural Resources
Colorado State University
Fort Collins, CO  80523
Phone: 970-491-7257
Hadley Wickham
# 
Hi Toby,

Have a look at the reshape package, http://had.co.nz.

Code something like the following should do what you want:

library(reshape)
dfm <- melt(data, id=1:3)
dfm <- rename(dfm, c(variable = "month")) # to make it more obvious

cast(dfm, ... ~ var)

Hadley
On 12/20/06, Toby Gass <tobygass at warnercnr.colostate.edu> wrote: