Restore trailing zeros

restore_zeros() takes a vector with values that might have lost trailing zeros, most likely from being registered as numeric. It turns each value into a string and adds trailing zeros until the mantissa hits some limit.

The default for that limit is the number of digits in the longest mantissa of the vector's values. The length of the integer part plays no role.

Don't rely on the default limit without checking: The original width could have been larger because the longest extant mantissa might itself have lost trailing zeros.

restore_zeros_df() is a variant for data frames. It wraps restore_zeros() and, by default, applies it to all columns that are coercible to numeric.

Usage

restore_zeros(
  x,
  width = NULL,
  sep_in = "\\.",
  sep_out = sep_in,
  sep = deprecated()
)

restore_zeros_df(
  data,
  cols = everything(),
  check_numeric_like = TRUE,
  check_decimals = FALSE,
  width = NULL,
  sep_in = "\\.",
  sep_out = NULL,
  sep = deprecated(),
  ...
)

Arguments

x: Numeric (or string coercible to numeric). Vector of numbers that might have lost trailing zeros.
width: Integer. Number of decimal places the mantissas should have, including the restored zeros. Default is NULL, in which case the number of characters in the longest mantissa will be used instead.
sep_in: Substring that separates the input's mantissa from its integer part. Default is "\\.", which renders a decimal point.
sep_out: Substring that will be returned in the output to separate the mantissa from the integer part. By default, sep_out is the same as sep_in.
sep: [Deprecated] Use sep_in, not sep. If sep is specified nonetheless, sep_in takes on sep's value.
data: Data frame or matrix. Only in restore_zeros_df(), and instead of x.
cols: Only in restore_zeros_df(). Select columns from data using tidyselect. Default is everything(), which selects all columns that pass the test of check_numeric_like.
check_numeric_like: Logical. Only in restore_zeros_df(). If TRUE (the default), the function will skip columns that are not numeric or coercible to numeric, as determined by is_numeric_like().
check_decimals: Logical. Only in restore_zeros_df(). If set to TRUE, the function will skip columns where no values have any decimal places. Default is FALSE.
...: Only in restore_zeros_df(). These dots must be empty.

Value

For restore_zeros(), a string vector. At least some of the strings will have newly restored zeros, unless (1) all input values had the same number of decimal places, and (2) width was not specified as a number greater than that single number of decimal places.
For restore_zeros_df(), a data frame.

Details

These functions exploit the fact that groups of summary values such as means or percentages are often reported to the same number of decimal places. If such a number is known but values were not entered as strings, trailing zeros will be lost. In this case, restore_zeros() or restore_zeros_df() will be helpful to prepare data for consistency testing functions such as grim_map() or grimmer_map().

Displaying decimal places

You might not see all decimal places of numeric values in a vector, and consequently wonder if restore_zeros(), when applied to the vector, adds too many zeros. That is because displayed numbers, unlike stored numbers, are often rounded.

For a vector x, you can count the characters of the longest mantissa from among its values like this:

x %>% decimal_places() %>% max()

Examples