如何在R中对向量进行分类以绘制饼图

Question

I want to categorize rivers dataset into “tiny” (<500), “short” (<1500), “medium” (<3000) and “long” (>=3000). I want to plot a pie chart that visualizes frequency of these four categories.

I tried:

 rivers[rivers >= 3000] = 'long'
 rivers[rivers >= 1500 & rivers < 3000] = 'meidum'
 rivers[rivers >= 500 & rivers < 1500]='short'
 rivers[rivers < 500] = 'tiny'

It seems the third command has no effect on data and they are the same as before!

table(rivers)
rivers
   500    505    524    525    529    538    540    545    560    570    600    605 
     2      1      1      2      1      1      1      1      1      1      3      1 
   610    618    620    625    630    652    671    680    696    710    720    730 
     1      1      1      1      1      1      1      1      1      1      2      1 
   735    760    780    800    840    850    870    890    900    906    981   long 
     2      1      1      1      1      1      1      1      2      1      1      1 
meidum   tiny 
    36     62 

What is wrong with my commands, and is it the right way to draw a pie chart for them?

Answer 1

The cut function and easily perform this task:

#random data
rivers<-runif(20, 0, 5000)

#break into desired groups and label
answer<-cut(rivers, breaks=c(0, 500, 1500, 3000, Inf), 
    labels=c("tiny", "short", "medium", "long"), right=FALSE) 

table(answer)
# tiny  short medium   long 
#    1     10      7      2 

Answer 2

You are running into this problem because you are trying to assign character values to an integer vector. If you work with a character vector instead, it should work:

> rivers_size <- as.character(rivers)
> rivers_size[rivers >= 3000] = 'long'
> rivers_size[rivers >= 1500 & rivers < 3000] = 'meidum'
> rivers_size[rivers >= 500 & rivers < 1500]='short'
> rivers_size[rivers < 500] = 'tiny'
> table(rivers_size)
rivers_size
  long meidum  short   tiny 
     1      5     53     82 
> pie(table(rivers_size))

Alternatively, the same thing can be accomplished using cut (as @Dave2e shows):

rivers <- cut(datasets::rivers,
              breaks = c(0, 500, 1500, 3000, Inf), 
              labels = c("tiny", "short", "medium", "long"),
              right = FALSE)
pie(table(rivers))

Answer 3

Here is another alternative using dplyr::case_when . It is more verbose than using cut but it is also easier generalize.

library("tidyverse")

set.seed(1234) # for reproducibility

# `case_when` vectorizes multiple `if-else` statements.
rivers <- sample.int(5000, size = 1000, replace = TRUE)
rivers <- case_when(
  rivers >= 3000 ~ "long",
  rivers >= 1500 ~ "medium",
  rivers >= 500  ~ "short",
  TRUE ~ "tiny"
)
table(rivers)
#> rivers
#>   long medium  short   tiny 
#>    406    303    199     92

^{Created on 2019-04-10 by the reprex package (v0.2.1)}