Grouping and filtering using `is.na` yields error.

[lschneiderbauer]

I am using the current development version of dbplyr.

Using filter with is.na yields in the following error:

library(dplyr, warn.conflicts = FALSE)
library(dbplyr, warn.conflicts = FALSE)

table <- tbl_lazy(tibble(col1 = c(1,2), col2 = c(1,2)), dbplyr::simulate_hana())

table %>%
  group_by(col2) %>%
  summarize(
    test = sum(col1, na.rm = TRUE),
    .groups = "drop"
  ) %>%
  filter(!is.na(test))
#> Error in eval_tidy(x, env = env): Objekt 'col1' nicht gefunden

^{Created on 2023-02-06 with reprex v2.0.2}

In contrast, using some other filter is working:

library(dplyr, warn.conflicts = FALSE)
library(dbplyr, warn.conflicts = FALSE)

table <- tbl_lazy(tibble(col1 = c(1,2), col2 = c(1,2)), dbplyr::simulate_hana())

table %>%
  group_by(col2) %>%
  summarize(
    test = sum(col1, na.rm = TRUE),
    .groups = "drop"
  ) %>%
  filter(test == "something else")
#> <SQL>
#> SELECT `col2`, SUM(`col1`) AS `test`
#> FROM `df`
#> GROUP BY `col2`
#> HAVING (SUM(`col1`) = 'something else')

^{Created on 2023-02-06 with reprex v2.0.2}

[mgirlich]

The issue only occurred when using a function on a summarised variable. This is why our tests didn't catch this. Similar to your second example they only used {summarised variable} >= 1.