convert dataframe from horizontal to vertical

I have data frame with values in this shape:

 col     x1   y1    x2   y2   x3  y3

 red     23  0.5    25  0.7   30  0.8

 blue    12  08     45  0.1   31  0.5

I need to convert this into two columns x and y:

col     x     y

I tried ‘pivot_longer’ but I can’t get the rigt pattern_sep or names_pattern

  df1<- pivot_longer(df,
                  col = !col,
                  names_sep = "(.)(.)"
          )   

>Solution :

You can use reshape in base R

reshape(
    df,
    direction = "long",
    idvar = "col",
    varying = -1,
    v.names = c("x", "y")
)

which gives

        col time  x   y
red.1   red    1 23 0.5
blue.1 blue    1 12 8.0
red.2   red    2 25 0.7
blue.2 blue    2 45 0.1
red.3   red    3 30 0.8
blue.3 blue    3 31 0.5

Or pivot_*

df %>%
    pivot_longer(
        cols = !col,
        names_pattern = "(.)(.)",
        names_to = c("name", "id")
    ) %>%
    pivot_wider(id_cols = c(col, id))

or much shorter (thanks for comments from @Ritchie Sacramento)

df %>%
    pivot_longer(-col, names_pattern = "(\\D+)(\\d+)", names_to = c(".value", "id"))

which gives

# A tibble: 6 × 4
  col   id        x     y
  <chr> <chr> <dbl> <dbl>
1 red   1        23   0.5
2 red   2        25   0.7
3 red   3        30   0.8
4 blue  1        12   8
5 blue  2        45   0.1
6 blue  3        31   0.5