Unable to extract Components from X13_ARIMA_SEATS model

[isachng93]

Dear community / @robjhyndman / @mitchelloharawild

I am attempting to perform seasonal adjustments on a grouped tsibble dataset using X13_ARIMA_SEATS. However, all went well until I attempted to extract the components from the mable.

This error does not occur when I attempt decomposition using STL or X11
The error also does not occur when I filter the mable to extract only a single row.

May I ask if anyone in the community is aware of a remedy to this issue?

Thanks much!

SCRIPT BELOW

#Global Functions#

singstat_to_tsibble_qtrly <- function(x) {x[["Data"]][["row"]] %>%
tidyr::unnest(cols = columns) %>%
dplyr::rename(Date = key) %>%
dplyr::rename(GroupId = rowText) %>%
dplyr::mutate(Date = tsibble::yearquarter(Date)) %>%
dplyr::mutate(value = as.numeric(value)) %>%
tsibble::tsibble(index = Date, key = GroupId, regular = TRUE) %>%
tsibble::group_by_key() %>%
dplyr::select(!c(seriesNo, uoM, footnote))}

#Crawl Data from Singstat#

VApW_Qtr_nonSA_set1_current_TS <- jsonlite::fromJSON("https://tablebuilder.singstat.gov.sg/api/table/tabledata/M015801?limit=2000") %>% singstat_to_tsibble_qtrly()
VApW_Qtr_nonSA_set2_current_TS <- jsonlite::fromJSON("https://tablebuilder.singstat.gov.sg/api/table/tabledata/M015801?offset=2000&limit=2000") %>% singstat_to_tsibble_qtrly()

#Combine dataset into 1 tsibble#

VApW_Qtr_nonSA_current_TS <- dplyr::full_join(VApW_Qtr_nonSA_set1_current_TS, VApW_Qtr_nonSA_set2_current_TS) %>%
dplyr::mutate(IndicatorType = "Performance",
uoM = "Value-added Per Worker",
Terms = "Nominal",
AdjStatus = "nonSA",
AdjMethod = "O")

#fit X13_ARIMA_SEATS model#

fit <- VApW_Qtr_nonSA_current_TS %>%
fabletools::model(X13 = feasts::X_13ARIMA_SEATS(value ~ seats(),
na.action = seasonal::na.x13,
defaults = "seasonal"))

#Extract Components#

fit %>% components()

---

Error occurs when I attempt to extract components from the X13_ARIMA_SEATS model.

rlang::last_trace()
<error/dplyr:::mutate_error>
Error in transmute():
ℹ In argument: cmp = map(.fit, components).
Caused by error in mutate():
ℹ In argument: seasonal = dcmp[, "adjustfac"].
Caused by error in dcmp[, "adjustfac"]:
! subscript out of bounds

Backtrace:
▆

1. ├─VApW_Qtr_SA_current_TS_test %>% components()
2. ├─generics::components(.)
3. ├─fabletools:::components.mdl_df(.)
4. │ ├─dplyr::transmute(...)
5. │ └─dplyr:::transmute.data.frame(...)
6. │ └─dplyr:::mutate_cols(.data, dots, by)
7. │ ├─base::withCallingHandlers(...)
8. │ └─dplyr:::mutate_col(dots[[i]], data, mask, new_columns)
9. │ └─mask$eval_all_mutate(quo)
10. │ └─dplyr (local) eval()
11. ├─fabletools:::map(.fit, components)
12. │ └─base::lapply(.x, .f, ...)
13. │ ├─generics (local) FUN(X[[i]], ...)
14. │ └─fabletools:::components.mdl_ts(X[[i]], ...)
15. │ ├─generics::components(object$fit, ...)
16. │ └─feasts:::components.feasts_x13arimaseats(object$fit, ...)
17. │ └─.data %>% ...
18. ├─dplyr::mutate(...)
19. └─dplyr:::mutate.data.frame(...)
20. └─dplyr:::mutate_cols(.data, dplyr_quosures(...), by)

21. ├─base::withCallingHandlers(...)
    

22. └─dplyr:::mutate_col(dots[[i]], data, mask, new_columns)
    

23.   └─mask$eval_all_mutate(quo)
    

24.     └─dplyr (local) eval()
    

---

[rogerlop]

Have the same issue. However, seats decomposition did work for most of the series but not for all of them. Has anyone found a way to get over this issue?