From d4190cc9ad15d30cb8b840f8a6df25c006d8009f Mon Sep 17 00:00:00 2001
From: Yaron Gvili <rtpsw@hotmail.com>
Date: Fri, 7 Oct 2022 15:00:10 -0400
Subject: [PATCH 1/2] ARROW-17964: [C++] Range data comparison for struct type
 may go out of bounds

---
 cpp/src/arrow/compare.cc | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/cpp/src/arrow/compare.cc b/cpp/src/arrow/compare.cc
index c5406ee583f..d6eb6756284 100644
--- a/cpp/src/arrow/compare.cc
+++ b/cpp/src/arrow/compare.cc
@@ -305,6 +305,11 @@ class RangeDataEqualsImpl {
   Status Visit(const StructType& type) {
     const int32_t num_fields = type.num_fields();
 
+    if (left_.child_data.size() != static_cast<size_t>(num_fields) ||
+        right_.child_data.size() != static_cast<size_t>(num_fields)) {
+      result_ = false;
+      return Status::OK();
+    }
     auto compare_runs = [&](int64_t i, int64_t length) -> bool {
       for (int32_t f = 0; f < num_fields; ++f) {
         RangeDataEqualsImpl impl(options_, floating_approximate_, *left_.child_data[f],

From 62fdbe4d6f98bfe9a156642ffd8c30e4ea98fb4d Mon Sep 17 00:00:00 2001
From: Yaron Gvili <rtpsw@hotmail.com>
Date: Thu, 13 Oct 2022 09:13:52 -0400
Subject: [PATCH 2/2] add tests

---
 cpp/src/arrow/compare.cc    |  3 +++
 cpp/src/arrow/datum_test.cc | 47 +++++++++++++++++++++++++++++++++++++
 2 files changed, 50 insertions(+)

diff --git a/cpp/src/arrow/compare.cc b/cpp/src/arrow/compare.cc
index d6eb6756284..951e64f3dad 100644
--- a/cpp/src/arrow/compare.cc
+++ b/cpp/src/arrow/compare.cc
@@ -870,6 +870,9 @@ Status PrintDiff(const Array& left, const Array& right, int64_t left_offset,
     return Status::OK();
   }
 
+  ARROW_RETURN_NOT_OK(left.ValidateFull());
+  ARROW_RETURN_NOT_OK(right.ValidateFull());
+
   if (!left.type()->Equals(right.type())) {
     *os << "# Array types differed: " << *left.type() << " vs " << *right.type()
         << std::endl;
diff --git a/cpp/src/arrow/datum_test.cc b/cpp/src/arrow/datum_test.cc
index 8f962962a21..0918b04a87d 100644
--- a/cpp/src/arrow/datum_test.cc
+++ b/cpp/src/arrow/datum_test.cc
@@ -18,10 +18,12 @@
 #include <memory>
 #include <string>
 
+#include <gtest/gtest-spi.h>
 #include <gtest/gtest.h>
 
 #include "arrow/array/array_base.h"
 #include "arrow/array/array_binary.h"
+#include "arrow/array/array_nested.h"
 #include "arrow/chunked_array.h"
 #include "arrow/datum.h"
 #include "arrow/record_batch.h"
@@ -149,4 +151,49 @@ TEST(Datum, TotalBufferSize) {
   ASSERT_EQ(4, tab_datum.TotalBufferSize());
 }
 
+TEST(Datum, Equality) {
+  AssertDatumsEqual(ArrayFromJSON(struct_({field("a", int32())}), "[[0], [1], [2]]"),
+                    ArrayFromJSON(struct_({field("a", int32())}), "[[0], [1], [2]]"));
+  EXPECT_FATAL_FAILURE(
+      AssertDatumsEqual(ArrayFromJSON(struct_({field("a", int32())}), "[[0], [1], [2]]"),
+                        ArrayFromJSON(struct_({field("a", int64())}), "[[0], [1], [2]]")),
+      "Array types differed");
+  AssertDatumsEqual(ArrayFromJSON(struct_({field("a", int32()), field("b", int64())}),
+                                  "[[0, 0], [1, 1], [2, 2]]"),
+                    ArrayFromJSON(struct_({field("a", int32()), field("b", int64())}),
+                                  "[[0, 0], [1, 1], [2, 2]]"));
+  EXPECT_FATAL_FAILURE(
+      AssertDatumsEqual(ArrayFromJSON(struct_({field("a", int32()), field("b", int32())}),
+                                      "[[0, 0], [1, 1], [2, 2]]"),
+                        ArrayFromJSON(struct_({field("a", int32())}), "[[0], [1], [2]]")),
+      "Array types differed");
+  EXPECT_FATAL_FAILURE(
+      AssertDatumsEqual(ArrayFromJSON(struct_({field("a", int32())}), "[[0], [1], [2]]"),
+                        ArrayFromJSON(struct_({field("a", int32()), field("b", int32())}),
+                                      "[[0, 0], [1, 1], [2, 2]]")),
+      "Array types differed");
+  EXPECT_FATAL_FAILURE(
+      AssertDatumsEqual(
+          ChunkedArrayFromJSON(struct_({field("a", int32()), field("b", int32())}),
+                               {"[[0, 0], [1, 1], [2, 2]]"}),
+          ChunkedArrayFromJSON(struct_({field("a", int32())}), {"[[0], [1], [2]]"})),
+      "Failed");
+  EXPECT_FATAL_FAILURE(
+      AssertDatumsEqual(
+          ChunkedArrayFromJSON(struct_({field("a", int32())}), {"[[0], [1], [2]]"}),
+          ChunkedArrayFromJSON(struct_({field("a", int32()), field("b", int32())}),
+                               {"[[0, 0], [1, 1], [2, 2]]"})),
+      "Failed");
+
+  EXPECT_FATAL_FAILURE(
+      AssertDatumsEqual(
+          ArrayFromJSON(struct_({field("a", int32())}), "[[0], [1], [2]]"),
+          std::make_shared<StructArray>(
+              struct_({field("a", int32())}), 3,
+              std::vector<std::shared_ptr<Array>>{ArrayFromJSON(int32(), "[0, 1, 2]"),
+                                                  ArrayFromJSON(int32(), "[0, 1, 2]")},
+              NULLPTR, 0, 0)),
+      "Failed");
+}
+
 }  // namespace arrow