I've run into an interesting problem, which I can not quite get around (as the solution seems to cause problems when I want to map and mutate nested datasets. When I use parse_date_time on a time series panel set like so...
df<-structure(list(date_strt = structure(c(18450, 18434, 18418, 18402,
18466, 18498, 18514, 18530, 18482, 18786, 18802, 18306, 18290,
18370, 18386), class = "Date"), item_id = c(47010, 47010, 47010,
47010, 47010, 47010, 47010, 47010, 47010, 47010, 47010, 47010,
47010, 47010, 47010), locn_id = c(74, 74, 74, 74, 74, 74, 74,
74, 74, 74, 74, 74, 74, 74, 74), sale_unit = c(63, 57, 43, 58,
65, 58, 189, 91, 52, 9, 10, 47, 22, 47, 67), rprc_unit_amnt = c(1.45,
1.45, 1.45, 1.45, 1.45, 1.45, 1.45, 1.45, 1.45, 1.45, 1.45, 1.45,
1.45, 1.45, 1.45), sale_unit_amnt = c(1.45, 1.45, 1.45, 1.45,
1.45, 1.45, 0.962380952380952, 1.29208791208791, 1.45, 1.30444444444444,
1.301, 1.45, 1.45, 1.45, 1.45), prmt_ind = c(0, 0, 0, 0, 0, 0,
1, 1, 0, 1, 1, 0, 0, 0, 0), cluster = c(7, 7, 7, 7, 7, 7, 7,
7, 7, 7, 7, 7, 7, 7, 7)), row.names = c(NA, -15L), class = c("tbl_df",
"tbl", "data.frame"))
This works perfectly fine (although it also seems to cause issues for later mapping of mutates), but, is slow(ish) for when scaling up to 4mill+ rows.
df %>%
unnest(cols = c(data)) %>%
mutate(date_strt = lubridate::parse_date_time(date_strt, c("ymd", "dmy")))
When this data is nested with an item grouping (df %>% mutate(item_grup_id = 6231) (I'm assuming you can repeat the same df with another grouping number and use this instead of a full repex of the nested df). Anyhow I get this error if I attempt to do this in a nested mutate
nested_df %>%
mutate(data = map(data, function(df){
df %>%
mutate(date_strt = lubridate::parse_date_time(date_strt, c("ymd", "dmy")))
}))
Error: Problem with `mutate()` column `data`.
i `data = map(...)`.
x no applicable method for 'mutate' applied to an object of class "Date"
If I do the full (slow) dataset parsing, and map a mutate later with a similar coding structure, I get pretty much the same error
nested_df %>%
mutate(data = map(data, function(df){
df %>%
mutate(unite(ITEMDECPRO, item_id, prmt_ind, remove = FALSE))
}))
Error: Problem with `mutate()` column `data`.
i `data = map(...)`.
x no applicable method for 'mutate' applied to an object of class "c('POSIXct', 'POSIXt')"
i The error occurred in row 1.
Does anyone have any ideas about how to do this without causing an error, and also why this occurs?
Thanks for reading