This should act something like do, but will append columns to existing data frame.

append_cols(.data, expression)

• expression is evaluated in environment where . refers to .data (same semantics as do())
• expression must return data.frame with either 1 or nrow rows

Q: Do I even need this? Is there a pattern in which do() with a single group (entire data frame) works?

Create function to allow renaming with either a function or a regex

Drop all rows with any NA in a subset of variables

Add a function to expand a formula in a given environment

Function that does relevel with a pattern