Ben Baumer

> However, modern best practice is to use tbl() directly on an DBIConnection. Not clear how much work this will be, but it is not insignificant.

``` > library(etl) > foo foo %>% etl_create() Could not find schema initialization script Show Traceback Rerun with Debug Error in eval(rlang::parse_expr(x)) : attempt to apply non-function ```

asynchronous operations?

1

Since `etl_extract()` and/or `etl_load()` may take a long time, can we dispatch them to run asynchronously? https://stackoverflow.com/questions/18819505/asynchronous-command-dispatch-in-interactive-r

The default method: Converts factors to characters Quotes all strings Converts all columns to strings Replaces NA with NULL

etl_transform.etl_cities() broken, but only on R CMD check()?

3

I'm stumped. ```{r} devtools::test() ``` works just fine. But it fails during ```{r} devtools::check() ``` ``` Status: 1 ERROR checking tests ... ERROR Running ‘testthat.R’ Running the tests in ‘tests/testthat.R’...

The most obvious candidate is `etl_transform()`, since those operations are local. Download speed is usually limited by the Internet connection speed, not the CPU. Also with database locking I'm not...