[bug] left and right joins should be swapped

[daikonradish]

Describe the bug

Haskell's function calls are as such:

fn arg1 arg2 arg3

The joins are such that:

-- This is *supposed* to preserve all keys in the left table.
-- Non matching keys are replaced with NULL.
D.leftJoin ["colname"] leftDf rightDf

-- This is *supposed* to preserve all keys in the right table.
-- Non matching keys are replaced with NULL.
D.rightJoin ["colname"] leftDf rightDf -- all colnames in rightDf preserved, non-matching in left replaced with NULL

However, as pointed out in #156, the behavior appears to be swapped.

Suggested Fix

1. Verify that the left and right joins are indeed swapped.
2. Modify the code in src/DataFrame/Operations/Join.hs to ensure that the correct join puts the left dataset on the left and the right dataset on the right
3. Add and update tests in tests/Operations/Join.hs
4. Update documentation, because now to use this in conjuntion with pipes, the way to express it is a little tricky: df |> (D.leftJoin other). Note: the documentation should be updated both dataframe/docs/cookbook.md and here.

This task is suitable for anyone who's just getting familiar with our codebases!