How to reorder rows with uncommon date order

[Ellenz]

I have a dataset from 1-Jan, how to reorder it so that 1-Feb is row 1, 2-Feb is row 2, etc. Jan-31 is row 365? I want to reorder the rows this way, so that I can plot the x-axis in this order. And maybe add another column day_order, which is from 1 to 365. Thanks for your help.

I cannot upload the sample data, so that the first few lines are like this:
DF
date value
Jan-1 0.1
Jan-2 0.5
Jan-3 1.2
Jan-4 3.4
...
Dec-31 1.9

[joels]

The details might depend on the actual structure of your data, including whether it's already sorted by date as implied in your question, whether a date can appear more than once, and whether it's a leap year, but here's a contrived example:

library(tidyverse)

# Fake data
date = seq(as.Date("2018-01-01"), as.Date("2018-12-31"), "1 day") %>% 
         format("%b-%d")

set.seed(2)
dat = tibble(date, value=rnorm(365))

# Ordering
dat = dat %>% 
  mutate(order=c((365 - 30):365, 1:(365 - 31)),
         day.order=1:365) %>% 
  arrange(order) %>% 
  mutate(date = factor(date, levels=unique(date))) %>% 
  select(-order) %>% 
  arrange(date)

dat

   date    value day.order
   <fct>   <dbl>     <int>
 1 Feb-01  0.319        32
 2 Feb-02  1.08         33
 3 Feb-03 -0.284        34
 4 Feb-04 -0.777        35
 5 Feb-05 -0.596        36
 6 Feb-06 -1.73         37
 7 Feb-07 -0.903        38
 8 Feb-08 -0.559        39
 9 Feb-09 -0.247        40
10 Feb-10 -0.384        41

[Yarnabrina]

Why not simply slice?

library(dplyr)
#> 
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#> 
#>     filter, lag
#> The following objects are masked from 'package:base':
#> 
#>     intersect, setdiff, setequal, union

set.seed(seed = 32982)

dataset <- tibble(date = format(x = seq.Date(from = as.Date(x = "2019/01/01"),
                                             to = as.Date(x = "2019/12/31"),
                                             length.out = 365),
                                format = "%b-%d"),
                  value = rnorm(n = 365))

dataset %>%
  mutate(day_order = 1:365) %>%
  slice((31 + 1):365, 1:31)
#> # A tibble: 365 x 3
#>    date     value day_order
#>    <chr>    <dbl>     <int>
#>  1 Feb-01  1.51          32
#>  2 Feb-02 -1.27          33
#>  3 Feb-03  1.41          34
#>  4 Feb-04 -0.279         35
#>  5 Feb-05  1.13          36
#>  6 Feb-06 -0.572         37
#>  7 Feb-07 -0.0690        38
#>  8 Feb-08  1.91          39
#>  9 Feb-09 -0.527         40
#> 10 Feb-10  1.25          41
#> # ... with 355 more rows

^{Created on 2019-06-13 by the reprex package (v0.3.0)}

[Ellenz]

Thanks for your responses.
Actually, I meant Feb-1 has the day_order of 1, Feb-2 has the day_order of 2, etc. So Jan-31 in the next year becomes the day_order of 365. There are just 365 days, but the 1st day starts from Feb-2, so it spans two years but just uses part of each year. How to redo the orders?