Question

In a typical spreadsheet I am able to create a formula that references a cell on any row.

So let us say I have the following data:

a=1:10
b=11:20
c=data.frame(a,b)

Which would yield this:

    a  b
1   1 11
2   2 12
3   3 13
4   4 14
5   5 15
6   6 16
7   7 17
8   8 18
9   9 19
10 10 20

So in a spreadsheet I can place in what would be cell c2 the absolute formula c$b2*c$a1 and get the number 12. Or maybe even something more complex like ifelse(c$a>5,c$b*c$a[-1]*2,c$b*c$a[-1] using relative references.

    a  b   c
1   1 11  NA
2   2 12  12
3   3 13  26
4   4 14  42
5   5 15  60
6   6 16  80
7   7 17 204
8   8 18 252
9   9 19 304
10 10 20 360

I know R is not a spreadsheet, but is there a package or function that would allow me to treat data frames in such a manner or is there something fundamentally wrong with that kind of functionality? I keep having these issues with functions that require the spanning of more than one record or row. Almost all the stuff I work with is a time-series if that helps at all. Is there a document anywhere that explains such functions? Maybe there is a ?topic I missed somewhere.

Answer 1

You were close, you need to complete the offset by excluding the first and last values, then wrap all that in ifelse:

a = 1:10
b = 11:20
d = data.frame(a,b)
d$c = ifelse(d$a > 5, c(NA, d$b[-1] * d$a[-nrow(d)]) * 2, c(NA, d$b[-1] * d$a[-nrow(d)]))

(I choose d as a variable name since c is a very commonly used function.)

For a tidier answer keep the main calculation out of the ifelse and use it only for the multiplier:

 d$c = ifelse(d$a > 5, 2, 1) * c(NA, d$b[-1] * d$a[-nrow(d)])