<lazyframe>$group_by()$agg()$collect() may hang

[eitsupi]

I noticed this example will hang.

as_polars_df(mtcars)$group_by("cyl", foo = TRUE)$agg()

This is not confirmed by Python Polars or current R Polars.
It may occur if some feature flag is missing, but I have not been able to identify it (Enabling several features did not solve the problem.).

[eitsupi]

It hangs.

as_polars_df(mtcars)$group_by(1:2, foo = TRUE)$agg()

It works.

as_polars_df(data.frame(a = 1:10))$group_by(1:2, foo = TRUE)$agg()

It works.

as_polars_df(data.frame(a = 1:14))$group_by(1:2, foo = TRUE)$agg()

It hangs.

as_polars_df(data.frame(a = 1:15))$group_by(1:2, foo = TRUE)$agg()

It works.

as_polars_df(mtcars[1:14, ])$group_by(1:2, foo = TRUE)$agg()

I don't know why, but it seems to hang if the original data frame has more than 15 rows???

[eitsupi]

Hmmm, it works.

as_polars_lf(mtcars)$group_by(1:2, foo = TRUE)$agg()$collect(streaming = TRUE)

It hangs.

as_polars_lf(mtcars)$group_by(1:2, foo = TRUE)$agg()$collect(streaming = FALSE)

[eitsupi]

I think we need to implement this part to cancel hanged queries.

neo-r-polars/src/rust/src/r_threads.rs

Lines 177 to 182 in 49ac30a

	// TODO: uncomment this block
	// let res_res = extendr_api::eval_string("Sys.sleep(0)");
	// if res_res.is_err() {
	// r_println!("R user interrupt");
	// return Err("interupt by user".into());
	// }