[R] Adding new column to data frame and filling some rows of it - classes

[Gabor Grothendieck]

Gorjanc Gregor <Gregor.Gorjanc <at> bfro.uni-lj.si> writes:


: Few days ago I was asking on a list about adding a column and filling
: it in some rows. I was satisfied, but one thing raised my attention. I
: will show itthrough an example:
: 
: tmp <- data.frame(y1=1:4, f1=factor(c("A", "B", "C", "D")))
: f2 <- factor(c("Z", "Y"))
: 
: # I would like to add f2 to tmp and I know that values in f2 fit to
: # first two values in tmp. I used
: 
: tmp$f2 <- NA
: tmp[1:2, "f2"] <- f2
: tmp
:   y1 f1 f2
: 1  1  A  2
: 2  2  B  1
: 3  3  C NA
: 4  4  D NA
: 
: # However tmp$f2 is not factor. How can I make it to be a factor? I tried
: # with 
: class(tmp$f2) <- "factor"
: 
: # but I get this
: tmp
:   y1 f1   f2
: 1  1  A NULL
: 2  2  B <NA>
: 3  3  C <NA>
: 4  4  D <NA>
: Warning message: 
: corrupt data frame: columns will be truncated or padded with NAs in: 
format.data.frame(x, digits =
: digits) 
: 
: # I tried the other approach
: tmp <- data.frame(y1=1:4, f1=factor(c("A", "B", "C", "D")))
: f2 <- factor(c("Z", "Y"))
: tmp$f2 <- f2
: tmp[3:4, "f2"] <- NA
: tmp
:   y1 f1   f2
: 1  1  A    Z
: 2  2  B    Y
: 3  3  C <NA>
: 4  4  D <NA>
: 
: class(tmp$f2)
: [1] "factor"
: 
: # And I have now the same class for tmp$f2 as in f2. Is this the
: # only way?


Here are a few possibilities:

R> # 1
R> 
R> tmp <- data.frame(y1=1:4, f1=factor(c("A", "B", "C", "D")))
R> f2 <- factor(c("Z", "Y"))
R> 
R> length(f2) <- nrow(tmp)
R> tmp$f2 <- f2
R> tmp
  y1 f1   f2
1  1  A    Z
2  2  B    Y
3  3  C <NA>
4  4  D <NA>
R> 
R> # 2
R> 
R> tmp <- data.frame(y1=1:4, f1=factor(c("A", "B", "C", "D")))
R> f2 <- factor(c("Z", "Y"))
R> 
R> tmp$f2 <- factor(NA, levels = levels(f2))
R> tmp[seq(along = f2),"f2"] <- f2
R> tmp
  y1 f1   f2
1  1  A    Z
2  2  B    Y
3  3  C <NA>
4  4  D <NA>
R> 
R> # 3
R> 
R> tmp <- data.frame(y1=1:4, f1=factor(c("A", "B", "C", "D")))
R> f2 <- factor(c("Z", "Y"))
R> 
R> merge(tmp, list(f2 = f2), by = 0, all.x = TRUE)
  Row.names y1 f1   f2
1         1  1  A    Z
2         2  2  B    Y
3         3  3  C <NA>
4         4  4  D <NA>


---------------

Here is just the input:


# 1

tmp <- data.frame(y1=1:4, f1=factor(c("A", "B", "C", "D")))
f2 <- factor(c("Z", "Y"))

length(f2) <- nrow(tmp)
tmp$f2 <- f2
tmp

# 2

tmp <- data.frame(y1=1:4, f1=factor(c("A", "B", "C", "D")))
f2 <- factor(c("Z", "Y"))   

tmp$f2 <- factor(NA, levels = levels(f2))
tmp[seq(along = f2),"f2"] <- f2
tmp

# 3

tmp <- data.frame(y1=1:4, f1=factor(c("A", "B", "C", "D")))
f2 <- factor(c("Z", "Y"))   

merge(tmp, list(f2 = f2), by = 0, all.x = TRUE)