Data wrangling / formatting package outsourced from scrutiny

[lhdjung]

If we outsource multiple new "backend packages", as proposed in ianhussey/ERROR#6, one of them should be dedicated to data wrangling and formatting. However, it would also be a core (i.e., user-facing) package because data wrangling is needed in both internal development code and analysis scripts.

The new package would be focused on the idea of "number-strings" as defined in the docs of is_numeric_like(): "strings that can be coerced to numeric without introducing new NAs". Such strings are needed because many scrutiny / errorverse functions care about trailing zeros. However, some data wrangling functions might be relevant to the errorverse for reasons other than number-strings, such as row_to_colnames().

Functions to include in the new package

To assess number-string-like properties of objects:

• is_numeric_like()
• decimal_places(), decimal_places_scalar()
• decimal_places_df()

Dedicated "data wrangling" functions:

• restore_zeros(), restore_zeros_df()
• split_by_parens()
• before_parens(), inside_parens(); these two need to go together with split_by_parens() in any case
• row_to_colnames()

Others?