Help with tsibble

#1

Hi,

I am learning time series and started with converting data into tsibble object. Data contains three variables: Date ( in the form of 2020 05) showing year and month; Brands (as A, B, C) and Sales ( as quantity)

I was trying to do the following:

df <- df%>%
  mutate(Date = yearmonth(Date)) %>%
  as_tsibble(key = `Brands`, index = `Date`)

But I get this error:
Error: A valid tsibble must have distinct rows identified by key and index. i Please use duplicates() to check the duplicated rows.

However, there are no duplicates in the data.

Thanks for your help!

#2

I'm skeptical that there are no duplicates in the data.
How would you prove this.
If instead of your code that fails to create the tsibble you did

df%>%
  mutate(Date = yearmonth(Date)) %>% {table(.$Date,.$Brands)}

what would be the result ?

1 Like
#8

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.

#4

could you be more precise/explicit about how your problem shows itself?
I just wrote code that looks like that, and it works fine on my data...

#3

Interesting....Thanks @nirgrahamuk! Yes, I do see duplicates now. I deleted a Location variable before converting into tsibble, thinking I am only interested in Brands as key and corresponding Sales forgetting about unique combinations where location and Brands combination kept it as unique rows earlier.

So, now that I have included it back in the game, how can I specify a combination of two variables as key for tsibble. In this case, Location and Brands. The below code doesn't seem to work.

as_tsibble(key = c(Location, Brands), index = Date)

#5

I just found 2 duplicates even with the combinations....probably that is the problem. I will update you after fixing these duplicates...hopefully it will work then.

Thank you so much on helping me with this!

#7

It does work now...Thanks @nirgrahamuk with the help on duplicates!