Create hourly dateseries with group by few columns-CodePudding

Hi I have a table df_in with 4 columns as shown below.I am getting error while trying to create dateseries on hourly basis with group by few columns.

Gender = c("Male","Female")
Sec = c("AA","BB")
min_date = c("12/1/2018","12/9/2018")
max_date = c("12/3/2018","12/10/2018")
    
df_in = data.frame(Gender,Sec,min_date,max_date)

for 1st row number of rows will be 72 (3 days * 24 hours) and for 2nd row 48 rows (2 days * 24 hours)

df_out will look like

I have tried with below code

df_out = df_in %>% group_by(Gender, Sec) %>% mutate (DateSeries = seq(as.POSIXct(min_date), as.POSIXct(max_date), by="hour"))

Thanks in advance.

CodePudding user response：

Does this produce what you need?

df_in %>% 
  group_by(Gender, Sec) %>% 
  expand(
    nesting(Gender,Sec), 
    DateSeries = seq(as.POSIXct(min_date), as.POSIXct(max_date), by="hour")
  )
# A tibble: 2,162 x 3
# Groups:   Gender, Sec [2]
   Gender Sec   DateSeries         
   <chr>  <chr> <dttm>             
 1 Female BB    0012-09-20 00:00:00
 2 Female BB    0012-09-20 01:00:00
 3 Female BB    0012-09-20 02:00:00
 4 Female BB    0012-09-20 03:00:00
 5 Female BB    0012-09-20 04:00:00
 6 Female BB    0012-09-20 05:00:00
 7 Female BB    0012-09-20 06:00:00
 8 Female BB    0012-09-20 07:00:00
 9 Female BB    0012-09-20 08:00:00
10 Female BB    0012-09-20 09:00:00
# ... with 2,152 more rows

Your first effort with mutate won't work because you're trying to add a 721 element vector to a data frame with only one row. expand is the function you need: it creates and "all possible combinations" version of its inputs.