Fix to_timestamps to support Z for %z format specifier

rapidsai · Apr 7, 2022 · 557114d · 557114d
1 parent fb03c8b
commit 557114d
Show file tree

Hide file tree

Showing 2 changed files with 13 additions and 8 deletions.
diff --git a/cpp/src/strings/convert/convert_datetime.cu b/cpp/src/strings/convert/convert_datetime.cu
@@ -298,12 +298,14 @@ struct parse_datetime {
         }
         case 'z': {
           // 'z' format is +hh:mm -- single sign char and 2 chars each for hour and minute
-          auto const sign     = *ptr == '-' ? 1 : -1;
-          auto const [hh, lh] = parse_int(ptr + 1, 2);
-          auto const [mm, lm] = parse_int(ptr + 3, 2);
-          // revert timezone back to UTC
-          timeparts.tz_minutes = sign * ((hh * 60) + mm);
-          bytes_read -= lh + lm;
+          if (item.length == 5) {
+            auto const sign     = *ptr == '-' ? 1 : -1;
+            auto const [hh, lh] = parse_int(ptr + 1, 2);
+            auto const [mm, lm] = parse_int(ptr + 3, 2);
+            // revert timezone back to UTC
+            timeparts.tz_minutes = sign * ((hh * 60) + mm);
+            bytes_read -= lh + lm;
+          }
           break;
         }
         case 'Z': break;  // skip
@@ -574,6 +576,8 @@ struct check_datetime_format {
             auto const cvm = check_value(ptr + 3, 2, 0, 59);
             result         = (*ptr == '-' || *ptr == '+') && cvh.first && cvm.first;
             bytes_read -= cvh.second + cvm.second;
+          } else if (item.length == 1) {
+            result = *ptr == 'Z';
           }
           break;
         }

diff --git a/cpp/tests/strings/datetime_tests.cpp b/cpp/tests/strings/datetime_tests.cpp
@@ -144,16 +144,17 @@ TEST_F(StringsDatetimeTest, ToTimestampTimezone)
                                              "2019-07-17 02:34:56-0300",
                                              "2019-03-20 12:34:56+1030",
                                              "2020-02-29 12:00:00-0500",
+                                             "2022-04-07 09:15:00Z",
                                              "1938-11-23 10:28:49+0700"};
   auto strings_view = cudf::strings_column_view(strings);
   auto results      = cudf::strings::to_timestamps(
     strings_view, cudf::data_type{cudf::type_id::TIMESTAMP_SECONDS}, "%Y-%m-%d %H:%M:%S%z");
   cudf::test::fixed_width_column_wrapper<cudf::timestamp_s, cudf::timestamp_s::rep> expected{
-    131243025, 1563341696, 1553047496, 1582995600, -981664271};
+    131243025, 1563341696, 1553047496, 1582995600, 1649322900, -981664271};
   CUDF_TEST_EXPECT_COLUMNS_EQUAL(*results, expected);
 
   results = cudf::strings::is_timestamp(strings_view, "%Y-%m-%d %H:%M:%S%z");
-  cudf::test::fixed_width_column_wrapper<bool> is_expected({1, 1, 1, 1, 1});
+  cudf::test::fixed_width_column_wrapper<bool> is_expected({1, 1, 1, 1, 1, 1});
   CUDF_TEST_EXPECT_COLUMNS_EQUAL(*results, is_expected);
 }