Port groupby() from Python

[ramhiser]

While this is an interesting function, dplyr handles this case well. Also, groupby may not translate easily to R because of the lack of dictionaries.

[ramhiser]

It turns out that Python's groupby is very similar to how R's tapply behaves. The main differences are that Python's groupby:

1. Iterates with a key-value pair, similar to enumerate
2. Requires the iterable to be sorted by the key beforehand.

Difference 2 is especially important because otherwise, the iterator needs to be consumed before each group's data (i.e., factor) can be seen and processed.

[ramhiser]

A couple of examples where groupby is useful:

1. Averaging over continuous blocks
2. Partial means of vector elements

The second entry is marked as a duplicate of the first, but groupby provides an alternative solution using iterators.

[ramhiser]

Also, code up an iterator version of Run Length Encoding from base::rle().

Example usage:

> x <- c(rep(6:10, 1:5), 6)
> x
 [1]  6  7  7  8  8  8  9  9  9  9 10 10 10 10 10  6
> rle(x)
Run Length Encoding
  lengths: int [1:6] 1 2 3 4 5 1
  values : num [1:6] 6 7 8 9 10 6