Error creating prophet object - Error in filter_imp(.data, quo)

I am trying to create a prophet object of my data frame by passing grouped data using the following code. Essentially, I would like my predictions at group level, therefore I subset the data frame and create prophet object for predict and then merge all groups together. However, I am not certain about the error I am receiving which seems to be from dplyr which is used by prophet library.

modelGroup <- function(x) { cdpModel <- prophet(x, changepoint.prior.scale = 0.5) cdpFuture <- make_future_dataframe(cdpModel, periods = 52, freq = "week) forecast <- predict(cdpModel, cdpFuture) return ......

Here some sample data

Week Group Count 16-01-01 A 10 17-01-01 B 15 15-01-01 C 20

I can’t seem to figure out the error, my function call subsets the data by group and passes a date and count column data frame to the function.

results <- DataTable[, lapply(.SD, modelGroup), by = DISTRICT]

Traceback:

`Error in filter_impl(.data, quo) : Result must have length 22923 not 9 17. call = filter_imp(.data, quo), cppstack = structure(list(file = “” line -1L, stack = c("usr/lib64/R/library/dplyr.so(Rcpp::exception::exception(char const*, bool)+0x73), "/usr/lib64/R/library/dplyr/libs/dplyr.so(void Rcpp::stop<int, long>(char const*, int const&, long const&)+0x55) 16. filter_imp(.data, quo) 15. filter.tbl_df(d=tvk_df(.data), …) . . . 11. dplyr::filter(., !is.na(y)) . . . 3. df %>% dplyr::filter(!is.na(y)) 2. fit.prophet(m, df, …)

1. prophet(df = dataframe)`

[kevalshah90]

This isn’t an issue anymore. The columns had to be renamed to ds and y before passing them to the function to create prophet object and running the model. Thanks for your time.

[kevalshah90]

Interesting @bletham. I verified that the dataset (x) being passed to the function always has 2 columns, weekly time stamp (ds) and y. Before the function call, the data frame is being subset with each group and only the selected 2 cols are passed to the function modelGroup.