feat: Better error message when user passed = by mistake in filter()

NEWS.md

@@ -11,6 +11,8 @@
 
 * Added `na.last` and `ties.method` arguments support to `rank()` (@Yousa-Mirage, #329).
 
+* Better error message in `filter()` when a condition uses `=` instead of `==` (#340).
+
 ## Bug fixes
 
 * Fix `NA` handling in `cummin()`, `cumprod()`, `cumsum()` (@Yousa-Mirage, #326).

R/filter.R

@@ -86,6 +86,24 @@ filter.polars_data_frame <- function(.data, ..., .by = NULL) {
   mo <- attributes(.data)$maintain_grp_order %||% FALSE
   is_grouped <- !is.null(grps)
 
+  # Check whether user passed "=" instead of "=="
+  dots <- enquos(...)
+  elems_with_assignment <- dots[!is.null(names(dots)) & nzchar(names(dots))]
+  if (length(elems_with_assignment) > 0) {
+    msg <- paste0(
+      names(elems_with_assignment)[1],
+      " == ",
+      quo_text(elems_with_assignment[[1]])
+    )
+    cli_abort(
+      c(
+        "!" = "We detected a named input.",
+        "i" = "This usually means that you've used {.code =} instead of {.code ==}.",
+        "i" = "Did you mean {.code {msg}}?"
+      )
+    )
+  }
+
   polars_exprs <- translate_dots(
     .data,
     ...,

tests/testthat/_snaps/filter-lazy.md

@@ -4,5 +4,27 @@
       current$collect()
     Condition
       Error in `filter()`:
-      ! Evaluation failed in `$filter()`.
+      ! We detected a named input.
+      i This usually means that you've used `=` instead of `==`.
+      i Did you mean `x == 1`?
+
+---
+
+    Code
+      current$collect()
+    Condition
+      Error in `filter()`:
+      ! We detected a named input.
+      i This usually means that you've used `=` instead of `==`.
+      i Did you mean `x == 1`?
+
+---
+
+    Code
+      current$collect()
+    Condition
+      Error in `filter()`:
+      ! We detected a named input.
+      i This usually means that you've used `=` instead of `==`.
+      i Did you mean `a == mpg`?
 

tests/testthat/_snaps/filter.md

@@ -1,8 +1,30 @@
 # error message when using =
 
     Code
-      filter(test_pl, chars = "a")
+      filter(test_pl, x = 1)
     Condition
       Error in `filter()`:
-      ! Evaluation failed in `$filter()`.
+      ! We detected a named input.
+      i This usually means that you've used `=` instead of `==`.
+      i Did you mean `x == 1`?
+
+---
+
+    Code
+      filter(test_pl, !is.na(y), x = 1)
+    Condition
+      Error in `filter()`:
+      ! We detected a named input.
+      i This usually means that you've used `=` instead of `==`.
+      i Did you mean `x == 1`?
+
+---
+
+    Code
+      f_pl(mpg)
+    Condition
+      Error in `filter()`:
+      ! We detected a named input.
+      i This usually means that you've used `=` instead of `==`.
+      i Did you mean `a == mpg`?
 

tests/testthat/test-filter-lazy.R

@@ -358,10 +358,36 @@ test_that("works with a local variable defined in a function", {
 })
 
 test_that("error message when using =", {
-  test_pl <- pl$LazyFrame(chars = letters[1:3])
+  test_df <- tibble(x = 1, y = 2)
+  test_pl <- as_polars_lf(test_df)
 
+  expect_both_error(
+    test_pl |> filter(x = 1),
+    test_df |> filter(x = 1)
+  )
+  expect_snapshot_lazy(
+    test_pl |> filter(x = 1),
+    error = TRUE
+  )
+
+  expect_both_error(
+    test_pl |> filter(!is.na(y), x = 1),
+    test_df |> filter(!is.na(y), x = 1)
+  )
+  expect_snapshot_lazy(
+    test_pl |> filter(!is.na(y), x = 1),
+    error = TRUE
+  )
+
+  f_pl <- function(x) {
+    filter(test_pl, "a" = {{ x }})
+  }
+  f_df <- function(x) {
+    filter(test_df, "a" = {{ x }})
+  }
+  expect_both_error(f_pl(mpg), f_df(mpg))
   expect_snapshot_lazy(
-    test_pl |> filter(chars = "a"),
+    f_pl(mpg),
     error = TRUE
   )
 })

tests/testthat/test-filter.R

@@ -354,10 +354,36 @@ test_that("works with a local variable defined in a function", {
 })
 
 test_that("error message when using =", {
-  test_pl <- pl$DataFrame(chars = letters[1:3])
+  test_df <- tibble(x = 1, y = 2)
+  test_pl <- as_polars_df(test_df)
 
+  expect_both_error(
+    test_pl |> filter(x = 1),
+    test_df |> filter(x = 1)
+  )
+  expect_snapshot(
+    test_pl |> filter(x = 1),
+    error = TRUE
+  )
+
+  expect_both_error(
+    test_pl |> filter(!is.na(y), x = 1),
+    test_df |> filter(!is.na(y), x = 1)
+  )
+  expect_snapshot(
+    test_pl |> filter(!is.na(y), x = 1),
+    error = TRUE
+  )
+
+  f_pl <- function(x) {
+    filter(test_pl, "a" = {{ x }})
+  }
+  f_df <- function(x) {
+    filter(test_df, "a" = {{ x }})
+  }
+  expect_both_error(f_pl(mpg), f_df(mpg))
   expect_snapshot(
-    test_pl |> filter(chars = "a"),
+    f_pl(mpg),
     error = TRUE
   )
 })