recode function

Recode or rearrange or reshape variables or values to new values

Given a set of numeric codes, change their values to different values given a mapping function. Also included are the ability to reorder columns or to convert wide sets of columns to long form

rearrange(x,pattern)   #reorder the variables
wide2long(x,width, cname=NULL, idname = NULL, idvalues=NULL ,pattern=NULL) 
recode(x, where, isvalue, newvalue)  #recode text values to numeric values

Arguments

• x: A matrix or data frame of numeric values
• where: The column numbers to fix
• isvalue: A vector of values to change
• newvalue: A vector of the new values
• pattern: column order of repeating patterns
• width: width of long format
• cname: Variable names of long format
• idname: Name of first column
• idvalues: Values to fill first column

Details

Three functions for basic recoding are included.

recode: Sometime, data are entered as levels in an incorrect order. Once converted to numeric values, this can lead to confusion. recoding of the data to the correct order is straightforward, if tedious.

rearrange: Another tedious problem is when the output of one function needs to be arranged for better data handling in subsequent function. Specify a pattern of choosing the new columns.

wide2long: And then, having rearranged the data, perhaps convert the file to long format.

Returns

The reordered data

Author(s)

William Revelle

Note

Although perhaps useful, the recode function is definitely ugly code. For smaller data sets, the results from char2numeric back to the original will not work. char2numeric works column wise and orders the data in each column.

See Also

mlArrange in the psych package for a more general version of wide2long

Examples

x <- matrix(1:120,ncol=12) 
new <- rearrange(x,pattern = c(1,4, 7,10))
new 
long <- wide2long(x,width=3,pattern=c(1,4, 7,10))  #rearrange and then make wide

temp <- bfi[1:100,1:5]
isvalue <- 1:6
newvalue <- psych::cs(one,two,three,four,five,six)
newtemp <- recode(temp,1:5,isvalue,newvalue)
newtemp  #characters
temp.num <- psych::char2numeric(newtemp) #convert to numeric
temp.num  #notice the numerical values have changed
new.temp.num <- recode(temp.num, 1:5, isvalue=c(3,6,5,2,1,4), newvalue=1:6)
#note that because char2numeric works column wise, this will fail for small sets