Question

And when I initialize the new array, would I need every value in this field to be initialized with an actual numerical value?

Answer 1

How you append the new field will depend on the object you have. Options are given below. If the length() of the variable you are trying to append does not equal the number of rows of the object you wish to append to then you will need to fill the new column/variable first with NA and then append your variable into the correct elements of the newly created object. For example, with

df <- data.frame(var1 = 1:5, var2 = LETTERS[1:5])

and we want to add a vector c(1,3,4) into rows 1, 2, and 5 of a third variable in df, then we could do this like so:

df$var3 <- NA ## note the single NA is recycled to the correct length

Then using subsetting to insert the new data into the correct elements

df$var3[c(1,2,5)] <- c(1,3,4)

which gives:

> df$var3[c(1,2,5)] <- c(1,3,4)
> df
  var1 var2 var3
1    1    A    1
2    2    B    3
3    3    C   NA
4    4    D   NA
5    5    E    4

If the variable you wish to append is of the correct length, then you don’t need to initialise anything. Use one of the approaches described below to achieve what you want in that case, depending on whether you have a data frame or a matrix.

Assuming a data frame

If you have a data frame then there are several ways. For example, using

df <- data.frame(var1 = 1:10, var2 = LETTERS[1:10])

We can do

> df$var3 <- letters[1:10]
> df
   var1 var2 var3
1     1    A    a
2     2    B    b
3     3    C    c
4     4    D    d
5     5    E    e
6     6    F    f
7     7    G    g
8     8    H    h
9     9    I    i
10   10    J    j

I prefer transform()

df <- transform(df, var3 = letters[1:10])

or the related within()

df <- within(df, var3 <- letters[1:10]) ## note diff syntax

and you can also do

df[, "var3"] <- letters[1:10]

or

df[, 3] <- letters[1:10]

but the latter one doesn’t produce a variable with the name "var3" but instead gets a default name.

You can also use cbind() to bind a column to a data frame (also works for matrices). Note that you have to either pass a data frame as the first argument or call the "data.frame" method directly otherwise you don’t get a data frame as the result:

df <- cbind(df, var3 = letters[1:10])
df2 <- cbind.data.frame(var3 = letters[1:10], df)
## note `df2` will have `var3` as the first column here

Assuming a matrix

If you have a matrix, then the options are reduced essentially to using cbind(). Using

mat <- matrix(1:20, ncol = 2)

you can do

mat <- cbind(mat, 1:10)

The other options don’t work.