Skip to contents

This is a speedier implementation of as.data.frame() but does not provide the same sort of checks. It should be used with caution.

Usage

quick_df(x = NULL)

empty_df()

quick_dfl(...)

Arguments

x

A list or NULL (see return)

...

Columns as tag = value (passed to list())

Value

A data.frame; if x is NULL a data.frame with 0 rows and 0 columns is returned (similar to calling data.frame() but faster). empty_df() returns a data.frame with 0 rows and 0 columns.

Examples

# unnamed will use make.names()
x <- list(1:10, letters[1:10])
quick_df(x)
#>    X1 X2
#> 1   1  a
#> 2   2  b
#> 3   3  c
#> 4   4  d
#> 5   5  e
#> 6   6  f
#> 7   7  g
#> 8   8  h
#> 9   9  i
#> 10 10  j

# named is preferred
names(x) <- c("numbers", "letters")
quick_df(x)
#>    numbers letters
#> 1        1       a
#> 2        2       b
#> 3        3       c
#> 4        4       d
#> 5        5       e
#> 6        6       f
#> 7        7       g
#> 8        8       h
#> 9        9       i
#> 10      10       j

# empty data.frame
empty_df() # or quick_df(NULL)
#> data frame with 0 columns and 0 rows