From 1904d1a9ff54343471998523816c9e0a00f46797 Mon Sep 17 00:00:00 2001
From: "Robert (Bobby) Evans" <bobby@apache.org>
Date: Wed, 1 Dec 2021 13:00:16 -0600
Subject: [PATCH] Fix overflow for min calculation in strings::from_timestamps
 (#9793)

This fixes #9790

When converting a timestamp to a String it is possible for the %M min calculation to overflow an int32_t part way through casting. This moves that result to be an int64_t which avoids the overflow issues.

Authors:
  - Robert (Bobby) Evans (https://github.com/revans2)

Approvers:
  - Bradley Dice (https://github.com/bdice)
  - David Wendt (https://github.com/davidwendt)

URL: https://github.com/rapidsai/cudf/pull/9793
---
 cpp/src/strings/convert/convert_datetime.cu | 4 ++--
 cpp/tests/strings/datetime_tests.cpp        | 8 +++++---
 2 files changed, 7 insertions(+), 5 deletions(-)

diff --git a/cpp/src/strings/convert/convert_datetime.cu b/cpp/src/strings/convert/convert_datetime.cu
index 51a6a796ba3..8d0c5704a7b 100644
--- a/cpp/src/strings/convert/convert_datetime.cu
+++ b/cpp/src/strings/convert/convert_datetime.cu
@@ -707,9 +707,9 @@ struct from_timestamp_base {
    *     scale( 61,60) ->  1
    * @endcode
    */
-  __device__ int32_t scale_time(int64_t time, int64_t base) const
+  __device__ int64_t scale_time(int64_t time, int64_t base) const
   {
-    return static_cast<int32_t>((time - ((time < 0) * (base - 1L))) / base);
+    return (time - ((time < 0) * (base - 1L))) / base;
   };
 
   __device__ time_components get_time_components(int64_t tstamp) const
diff --git a/cpp/tests/strings/datetime_tests.cpp b/cpp/tests/strings/datetime_tests.cpp
index 4543607614f..9a01d5dd041 100644
--- a/cpp/tests/strings/datetime_tests.cpp
+++ b/cpp/tests/strings/datetime_tests.cpp
@@ -311,13 +311,14 @@ TEST_F(StringsDatetimeTest, FromTimestampAmPm)
 TEST_F(StringsDatetimeTest, FromTimestampMillisecond)
 {
   cudf::test::fixed_width_column_wrapper<cudf::timestamp_ms, cudf::timestamp_ms::rep> timestamps_ms{
-    1530705600123, 1582934461007, 1451430122421, 1318302183999, -6106017600047};
+    1530705600123, 1582934461007, 1451430122421, 1318302183999, -6106017600047, 128849018880000};
   auto results = cudf::strings::from_timestamps(timestamps_ms, "%Y-%m-%d %H:%M:%S.%3f");
   cudf::test::strings_column_wrapper expected_ms{"2018-07-04 12:00:00.123",
                                                  "2020-02-29 00:01:01.007",
                                                  "2015-12-29 23:02:02.421",
                                                  "2011-10-11 03:03:03.999",
-                                                 "1776-07-04 11:59:59.953"};
+                                                 "1776-07-04 11:59:59.953",
+                                                 "6053-01-23 02:08:00.000"};
   CUDF_TEST_EXPECT_COLUMNS_EQUAL(*results, expected_ms);
 
   results = cudf::strings::from_timestamps(timestamps_ms, "%Y-%m-%d %H:%M:%S.%f");
@@ -325,7 +326,8 @@ TEST_F(StringsDatetimeTest, FromTimestampMillisecond)
                                                     "2020-02-29 00:01:01.007000",
                                                     "2015-12-29 23:02:02.421000",
                                                     "2011-10-11 03:03:03.999000",
-                                                    "1776-07-04 11:59:59.953000"};
+                                                    "1776-07-04 11:59:59.953000",
+                                                    "6053-01-23 02:08:00.000000"};
   CUDF_TEST_EXPECT_COLUMNS_EQUAL(*results, expected_ms_6f);
 
   cudf::test::fixed_width_column_wrapper<cudf::timestamp_ns, cudf::timestamp_ns::rep> timestamps_ns{