scrub {psych} | R Documentation |
A tedious part of data analysis is addressing the problem of miscoded data that need to be converted to NA or some other value. For a given data.frame or matrix, scrub will set all values of columns from=from to to=to that are less than a set (vector) of min values or more than a vector of max values to NA. Can also be used to do basic recoding of data for all values=isvalue to newvalue. Will also recode continuus variables into fewer categories. Will convert Nan, -Inf and Inf to NA
The length of the where, isvalue, and newvalues must either match, or be 1.
scrub(x, where, min, max,isvalue,newvalue, cuts=NULL)
x |
a data frame or matrix |
where |
The variables to examine. (Can be by name or by column number) |
min |
a vector of minimum values that are acceptable |
max |
a vector of maximum values that are acceptable |
isvalue |
a vector of values to be converted to newvalue (one per variable) |
newvalue |
a vector of values to replace those that match isvalue |
cuts |
The lower,and upper boundaries for recoding |
Solves a tedious problem that can be done directly but that is sometimes awkward. Will either replace specified values with NA or will recode to values within a range.
The corrected data frame.
William Revelle
reverse.code
, rescale
for other simple utilities.
data(attitude)
x <- scrub(attitude,isvalue=55) #make all occurrences of 55 NA
x1 <- scrub(attitude, where=c(4,5,6), isvalue =c(30,40,50),
newvalue = c(930,940,950)) #will do this for the 4th, 5th, and 6th variables
x2 <- scrub(attitude, where=c(4,4,4), isvalue =c(30,40,50),
newvalue = c(930,940,950)) #will just do it for the 4th column
new <- scrub(attitude,1:3,cuts= c(10,40,50,60,100)) #change many values to fewer
#get rid of a complicated set of cases and replace with missing values
y <- scrub(attitude,where=2:4,min=c(20,30,40),max= c(120,110,100),isvalue= c(32,43,54))
y1 <- scrub(attitude,where="learning",isvalue=55,newvalue=999) #change a column by name
y2 <- scrub(attitude,where="learning",min=45,newvalue=999) #change a column by name
y3 <- scrub(attitude,where="learning",isvalue=c(45,48),
newvalue=999) #change a column by name look for multiple values in that column
y4 <- scrub(attitude,where="learning",isvalue=c(45,48),
newvalue= c(999,-999)) #change values in one column to one of two different things