Question

I want to transpose data sets similar to my.data below and then sum the rows.

my.data <- "landuse units year county.a  county.b  county.c  county.d 
            apple   acres 2010     0         2         4         6 
            pear    acres 2010    10        20        30        40
            peach   acres 2010   500       400       300       200"

my.data2 <- read.table(textConnection(my.data), header = T)
my.data2

The desired output is:

 counties all.fruit
 county.a       510
 county.b       422
 county.c       334
 county.d       246

I can do this with the code below. However, the following code seems like it must be enormous overkill. I am hoping there is a much simpler solution.

# transpose the data set

tmy.data2 <- t(my.data2)
tmy.data2 <- as.data.frame(tmy.data2)

# assign row names to the data set

my.rows <- row.names(tmy.data2)

transposed.data <- cbind(my.rows, tmy.data2)
transposed.data

# extract numbers to obtain row sums

fruit.data <- as.data.frame(transposed.data[4:dim(transposed.data)[1], 2:dim(transposed.data)[2]])

fruit.data2 <- as.matrix(fruit.data)

fruit.data3 <- matrix(as.numeric(fruit.data2), nrow=( dim(fruit.data2)[1] ),  byrow=F)

# sum fruit by county

all.fruit <- rowSums(fruit.data3, na.rm=T)

# create row names for summed fruit data

counties <- my.rows[4:length(my.rows)]

almost.final.data <- cbind(counties, all.fruit)

really.final.data <- as.data.frame(almost.final.data)

really.final.data[,2] <- as.numeric(as.character(really.final.data[,2]))
really.final.data

str(really.final.data)

Thank you for any suggestions. I can use the code above, but view this request as an opportunity to greatly improve my programming.

Answer 1

I would just subset the "county" columns, sum them, and create a data.frame using the results:

out <- colSums(my.data2[,grepl("county",colnames(my.data2))])
out2 <- data.frame(counties=names(out), all.fruit=out,
          row.names=NULL, stringsAsFactors=FALSE)