Clever Geek Handbook

How to group the results and count each group written to the matrix in R?

Here's the table:

Month    D1       D2      D3     D4 
1.      11149   10488    5593    3073
1.      6678    11073    789     10009
2.      2322    10899    3493    21
3.      5839    11563    4367    9987

I want to split all of the above content (4 columns of distance by 4 rows of the month) in 3 groups and have the counts of each group written in the matrix as such:

Month        D<=700    700<D<1000   D>=1000
1.           counts       counts    ....
2.           ...          
3.           ....

What's the fastest way to do this?

Thank!

+3


source to share


2 answers


Solution using package data.table

:

library(data.table)
library(magrittr)

setDT(dt)[, cut(colSums(.SD),breaks=c(0,700,1000,max(colSums(.SD)))) %>%
              table %>%
              as.list
          , Month]

#   Month (0,700] (700,1e+03] (1e+03,2.16e+04]
#1:     1       0           0                4
#2:     2       1           0                3
#3:     3       0           0                4


Data:

dt = structure(list(Month = c(1, 1, 2, 3), D1 = c(11149L, 6678L, 2322L, 
5839L), D2 = c(10488L, 11073L, 10899L, 11563L), D3 = c(5593L, 
789L, 3493L, 4367L), D4 = c(3073L, 10009L, 21L, 9987L)), .Names = c("Month", 
"D1", "D2", "D3", "D4"), class = "data.frame", row.names = c(NA, 
-4L))
+2


source


It doesn't seem like the fastest way, but you can use melt()

and cast()

from library(reshape)

: (assuming that d

's your original data.frame

)



library(reshape)
M <- melt(d,id.vars="month")
M$class <- cut(M$value,breaks=c(0,700,1000,max(M$value)))
C <- cast(M,month~class)
+1


source






More articles:


All Articles