Data types in a frame

A categorical variable is related to what is sometimes called an "enum" in other languages. R calls these "factors" and represents them efficiently as integers. For code designed for categorical variables, we need to tell R that this column of a data frame should be treated as a "factor."

Here is a column of strings:



In [60]:

    
stringcolors = c("red","red","blue","green","red")



In [61]:

    
stringcolors









    Out[61]:





	'red'
	'red'
	'blue'
	'green'
	'red'

Let's convert it to factors.



In [62]:

    
colors = factor(stringcolors)



In [63]:

    
colors









    Out[63]:





	red
	red
	blue
	green
	red



In [64]:

    
is.factor(colors)









    Out[64]:




TRUE



In [65]:

    
sizes = factor(c("S","M","M","L","S"))



In [66]:

    
prices = factor(c(19.99, 12.99, 9.99,12.99,19.99))



In [67]:

    
prices









    Out[67]:





	19.99
	12.99
	9.99
	12.99
	19.99



In [68]:

    
clothes = data.frame(colors,sizes,prices)



In [69]:

    
clothes









    Out[69]:





colors sizes prices

	1 red S 19.99
	2 red M 12.99
	3 blue M 9.99
	4 green L 12.99
	5 red S 19.99



In [70]:

    
summary(clothes)









    Out[70]:





   colors  sizes   prices 
 blue :1   L:1   9.99 :1  
 green:1   M:2   12.99:2  
 red  :3   S:2   19.99:2



In [71]:

    
xtabs(~colors+sizes, data=clothes)









    Out[71]:





       sizes
colors  L M S
  blue  0 1 0
  green 1 0 0
  red   0 1 2

But the sizes are in the wrong order; R has no way to know what it is. To tell R to treat a variable as ordinal, we do this:



In [72]:

    
sizes = factor(sizes,levels=c("S","M","L"), ordered=TRUE)



In [74]:

    
clothes = data.frame(colors,sizes,prices)



In [75]:

    
xtabs(~colors+sizes, data=clothes)









    Out[75]:





       sizes
colors  S M L
  blue  0 1 0
  green 0 0 1
  red   2 1 0

Now the sizes have been ordered S<M<L



In [ ]: