Rethrow "overflow" join error with a better message (#6956)

* Rethrow join explosion error

* NEWS bullet

* R -> dplyr

* Simplify message

NEWS.md

@@ -1,5 +1,8 @@
 # dplyr (development version)
 
+* `left_join()` and friends now return a specialized error message if they
+  detect that your join would return more rows than dplyr can handle (#6912).
+
 * `slice_*()` now throw the correct error if you forget to name `n` while also
   prefixing the call with `dplyr::` (#6946).
 

R/join-rows.R

@@ -107,6 +107,9 @@ dplyr_locate_matches <- function(needles,
     vctrs_error_incompatible_type = function(cnd) {
       abort("`join_cast_common()` should have handled this.", .internal = TRUE)
     },
+    vctrs_error_matches_overflow = function(cnd) {
+      rethrow_error_join_matches_overflow(cnd, error_call)
+    },
     vctrs_error_matches_nothing = function(cnd) {
       rethrow_error_join_matches_nothing(cnd, error_call)
     },
@@ -137,6 +140,26 @@ dplyr_locate_matches <- function(needles,
   )
 }
 
+rethrow_error_join_matches_overflow <- function(cnd, call) {
+  size <- cnd$size
+
+  stop_join(
+    message = c(
+      "This join would result in more rows than dplyr can handle.",
+      i = glue(
+        "{size} rows would be returned. ",
+        "2147483647 rows is the maximum number allowed."
+      ),
+      i = paste0(
+        "Double check your join keys. This error commonly occurs due to a ",
+        "missing join key, or an improperly specified join condition."
+      )
+    ),
+    class = "dplyr_error_join_matches_overflow",
+    call = call
+  )
+}
+
 rethrow_error_join_matches_nothing <- function(cnd, call) {
   i <- cnd$i
 

tests/testthat/_snaps/join-rows.md

@@ -344,6 +344,16 @@
       ! `relationship` must be one of "one-to-one", "one-to-many", "many-to-one", or "many-to-many", not "warn-many-to-many".
       i Did you mean "many-to-many"?
 
+# join_rows() rethrows overflow error nicely (#6912)
+
+    Code
+      join_rows(df, df, condition = ">=")
+    Condition
+      Error:
+      ! This join would result in more rows than dplyr can handle.
+      i 50000005000000 rows would be returned. 2147483647 rows is the maximum number allowed.
+      i Double check your join keys. This error commonly occurs due to a missing join key, or an improperly specified join condition.
+
 # `multiple = NULL` is deprecated and results in `'all'` (#6731)
 
     Code

tests/testthat/test-join-rows.R

@@ -410,6 +410,19 @@ test_that("join_rows() validates `relationship`", {
   })
 })
 
+test_that("join_rows() rethrows overflow error nicely (#6912)", {
+  skip_on_cran()
+  # Windows 32-bit doesn't support long vectors of this size, and the
+  # intermediate `r_ssize` will be too large
+  skip_if(.Machine$sizeof.pointer < 8L, message = "No long vector support")
+
+  df <- tibble(x = 1:1e7)
+
+  expect_snapshot(error = TRUE, {
+    join_rows(df, df, condition = ">=")
+  })
+})
+
 # Deprecated behavior ----------------------------------------------------------
 
 test_that("`multiple = NULL` is deprecated and results in `'all'` (#6731)", {