Question

I have a large data frame that I am attempting to run a for loop, but need to get all the different observations first. The process that I am currently trying to approach is this:

1. Create a list/vector/data frame of the observation
2. Use each observation in a for loop

Here is how I tackled it:

Generate some data:

x <- gl(12,5)

The result I am looking for is a list of all the different observations:

1 2 3 4 5 6 7 8 9 10 11 12

I could create a data frame, then drop the number columns like this:

list <- data.frame(table(x))

list[,2] <- NULL

print(list)

From this list/vector/dataframe, I would run through the for loop like so:

for (i in list) print(i)

There must be a more efficient way to get the list for the for loop. Can someone suggest one?

Answer 1

Several options include:

R> unique(x)
 [1] 1  2  3  4  5  6  7  8  9  10 11 12
Levels: 1 2 3 4 5 6 7 8 9 10 11 12
R> levels(x)
 [1] "1"  "2"  "3"  "4"  "5"  "6"  "7"  "8"  "9"  "10" "11" "12"
R> as.numeric(levels(x))
 [1]  1  2  3  4  5  6  7  8  9 10 11 12

The levels() ones are most logical and appropriate if you want the groupings.

Now whether you really want this is another question. See ?split which can split up a data frame on the basis of a factor (which you can lapply() over), or ?aggregate among others. There is also the plyr package which seems very popular for this sort of thing.