Joran Elias
Joran Elias
--- My BigQuery instance only appears to acknowledge a type of FLOAT64 while bigrquery appears to, when uploading data, translate SQL casting to FLOAT. I'm on R 4.1.1, Windows 10...
Looks at new_data for the number of complete rows to determine the chunk size, then generates the predicted values in chunks. It leads to some significant performance improvements in cases...
I suspect it should be possible to improve the performance for generating forecasts from recursive models by generating them in chunks rather than one row at a time. As an...
It might potentially be convenient, particularly when modeling nested time series that have somewhat different scales, to be able to specify the logistic_cap/logistic_floor in a relative manner. i.e. to be...
The most recent version of recipes 1.0.0 has deprecated the function `terms_select()` in favor of `recipes_eval_select()`. Everything still works, obviously, but it generates a lot of warnings.
I know we weren't sure whether these belonged in a separate package or not, but I went ahead and just made them since it was pretty easy. No worries if...
Some common data wrangling tools that are used in dplyr-like pipe chains live in tidyr, rather than dplyr. Some would be fairly simple to create "nested" versions, since in general...
I was profiling some code that calls `padr::pad()` and found that `span_all_groups()` was using a significant chunk of the total time. This proposed change converts `span_all_groups()` to use `purrr::map2` and...
This is an extremely minimal first pass. My aim here was to make creating & scheduling a build work when the yaml specifies a private worker pool by setting the...
I apologize for asking here, but I've simply hit a wall trying to figure this out from Google's documentation. I'm trying to submit a build using googleCloudRunner to a private...