R - Assign value to multiple cell based on a combination of values

Question

I have the following data.frame , where the multiple X columns (1,2,3... N) are blank:

df1 <- data.frame( name = c("A","B","C"), 
                   X1 = c("","", ""), 
                   Y1 = c("aa","bb","cc"), 
                   Z1 = c("AA","BB","CC"),
                   X2 = c("","", ""), 
                   Y2 = c("dd","",""),
                   Z2 = c("AA","",""),
                   X3 = c("","", ""), 
                   Y3 = c("","","ee"), 
                   Z3 = c("","","CC"))

Another data.frame contains the value that should be assigned to the X columns accordingly to the combination of values observed in the Ys and Zz columns:

df2 <- data.frame( Y = c("aa","bb","cc","dd","ee"), 
                   Z = c("AA","BB","CC","AA","CC"),
                   X = c (1,2,3,4,5))

How could I assign the values of X in df1 based on the information I have on df2, So I can get df3?:

df3 <- data.frame( name = c("A","B","C"), 
                   X1 = c("1","2", "3"), 
                   Y1 = c("aa","bb","cc"), 
                   Z1 = c("AA","BB","CC"),
                   X2 = c("4","", ""), 
                   Y2 = c("dd","",""),
                   Z2 = c("AA","",""),
                   X3 = c("","", "5"), 
                   Y3 = c("","","ee"), 
                   Z3 = c("","","CC"))`

Please note that in my real database each name may contain, ,but not necessarily does, several columns (for example, X1,Y1,Z1... X10,Y10,Z10 ).

Answer 1

This strategy reshapes your data from a wide format to a long format, does the merge, then reshapes everything back.

# go from wide to long
x1 <- reshape(df1, 
    varying=Map(function(x) paste0(x, 1:3), c("X","Y","Z")),
    v.names=c("X","Y","Z"),
    idvar="name",    
    timevar="time",
    direction="long")

x2 <- merge(subset(x1, select=-X), df2, by=c("Y","Z"), all.x=T)
# replace NA values with blanks
x2[is.na(x2$X),"X"] <- ""

# go back to wide
x3 <- reshape(x2,idvar="name",direction="wide", sep="")

and x3 is then

  name Y1 Z1 X1 Y2 Z2 X2 Y3 Z3 X3
1    A aa AA  1 dd AA  4         
2    B bb BB  2                  
3    C cc CC  3          ee CC  5

Here you get the columns in a slightly different order but you can easily fix after the fact if necessary.

You can see there is one place i hard-coded 1:3 . If you have more repetitions of columns, you can adjust that vector.