Upmerge to the latest CUDF and fix compile errors (NVIDIA#1154)

Signed-off-by: Robert (Bobby) Evans <[email protected]>
mythrocks · May 19, 2023 · a6985de · a6985de
1 parent 318bdbd
commit a6985de
Show file tree

Hide file tree

Showing 5 changed files with 30 additions and 12 deletions.
diff --git a/src/main/cpp/src/cast_string.cu b/src/main/cpp/src/cast_string.cu
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2022, NVIDIA CORPORATION.
+ * Copyright (c) 2022-2023, NVIDIA CORPORATION.
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
@@ -21,6 +21,7 @@
 
 #include <cudf/column/column.hpp>
 #include <cudf/column/column_device_view.cuh>
+#include <cudf/detail/null_mask.hpp>
 #include <cudf/detail/utilities/cuda.cuh>
 #include <cudf/detail/utilities/integer_utils.hpp>
 #include <cudf/null_mask.hpp>
@@ -652,7 +653,8 @@ struct string_to_integer_impl {
                                      rmm::mr::device_memory_resource* mr)
   {
     if (string_col.size() == 0) {
-      return std::make_unique<column>(data_type{type_to_id<T>()}, 0, rmm::device_buffer{});
+      return std::make_unique<column>(data_type{type_to_id<T>()}, 0, rmm::device_buffer{},
+              rmm::device_buffer{}, 0);
     }
 
     rmm::device_uvector<T> data(string_col.size(), stream, mr);
@@ -672,8 +674,10 @@ struct string_to_integer_impl {
       ansi_mode,
       strip);
 
+    auto null_count = cudf::detail::null_count(null_mask.data(), 0, string_col.size(), stream);
+
     auto col = std::make_unique<column>(
-      data_type{type_to_id<T>()}, string_col.size(), data.release(), null_mask.release());
+      data_type{type_to_id<T>()}, string_col.size(), data.release(), null_mask.release(), null_count);
 
     if (ansi_mode) { validate_ansi_column(col->view(), string_col, stream); }
 
@@ -737,8 +741,11 @@ struct string_to_decimal_impl {
       precision,
       strip);
 
+    auto null_count = cudf::detail::null_count(null_mask.data(), 0, string_col.size(), stream);
+
     auto col =
-      std::make_unique<column>(dtype, string_col.size(), data.release(), null_mask.release());
+      std::make_unique<column>(dtype, string_col.size(), data.release(),
+              null_mask.release(), null_count);
 
     if (ansi_mode) { validate_ansi_column(col->view(), string_col, stream); }
 
@@ -818,7 +825,9 @@ std::unique_ptr<column> string_to_decimal(int32_t precision,
       CUDF_FAIL("Unable to support decimal with precision " + std::to_string(precision));
   }();
 
-  if (string_col.size() == 0) { return std::make_unique<column>(dtype, 0, rmm::device_buffer{}); }
+  if (string_col.size() == 0) {
+    return std::make_unique<column>(dtype, 0, rmm::device_buffer{}, rmm::device_buffer{}, 0);
+  }
 
   return type_dispatcher(
     dtype, detail::string_to_decimal_impl{}, dtype, precision, string_col, ansi_mode, strip, stream, mr);

diff --git a/src/main/cpp/src/map_utils.cu b/src/main/cpp/src/map_utils.cu
@@ -628,8 +628,11 @@ std::unique_ptr<cudf::column> from_json(cudf::column_view const &input,
   auto structs_col = cudf::make_structs_column(num_pairs, std::move(out_keys_vals), 0,
                                                rmm::device_buffer{}, stream, mr);
 
+  auto offsets = std::make_unique<cudf::column>(std::move(list_offsets), 
+              rmm::device_buffer{}, 0);
+
   return cudf::make_lists_column(
-      input.size(), std::make_unique<cudf::column>(std::move(list_offsets)), std::move(structs_col),
+      input.size(), std::move(offsets), std::move(structs_col),
       input.null_count(), cudf::detail::copy_bitmask(input, stream, mr), stream, mr);
 }
 

diff --git a/src/main/cpp/src/row_conversion.cu b/src/main/cpp/src/row_conversion.cu
@@ -1886,10 +1886,13 @@ std::vector<std::unique_ptr<column>> convert_to_rows(
                    auto const offset_count = batch_info.row_batches[batch].row_offsets.size();
                    auto offsets = std::make_unique<column>(
                        data_type{type_id::INT32}, (size_type)offset_count,
-                       batch_info.row_batches[batch].row_offsets.release());
+                       batch_info.row_batches[batch].row_offsets.release(),
+                       rmm::device_buffer{}, 0);
                    auto data = std::make_unique<column>(data_type{type_id::INT8},
                                                         batch_info.row_batches[batch].num_bytes,
-                                                        std::move(output_buffers[batch]));
+                                                        std::move(output_buffers[batch]),
+                                                        rmm::device_buffer{},
+                                                        0);
 
                    return make_lists_column(
                        batch_info.row_batches[batch].row_count, std::move(offsets), std::move(data),

diff --git a/src/main/cpp/tests/cast_string.cpp b/src/main/cpp/tests/cast_string.cpp
@@ -240,7 +240,8 @@ TYPED_TEST(StringToIntegerTests, Overflow)
 
 TYPED_TEST(StringToIntegerTests, Empty)
 {
-  auto empty = std::make_unique<column>(data_type{type_id::STRING}, 0, rmm::device_buffer{});
+  auto empty = std::make_unique<column>(data_type{type_id::STRING}, 0, rmm::device_buffer{},
+          rmm::device_buffer{}, 0);
 
   auto result = spark_rapids_jni::string_to_integer(data_type{type_to_id<TypeParam>()},
                                                     strings_column_view{empty->view()},
@@ -541,7 +542,8 @@ TEST_F(StringToDecimalTests, Edges)
 
 TEST_F(StringToDecimalTests, Empty)
 {
-  auto empty = std::make_unique<column>(data_type{type_id::STRING}, 0, rmm::device_buffer{});
+  auto empty = std::make_unique<column>(data_type{type_id::STRING}, 0, rmm::device_buffer{},
+          rmm::device_buffer{}, 0);
 
   auto const result = spark_rapids_jni::string_to_decimal(
     8, 2, strings_column_view{empty->view()}, false, true, cudf::get_default_stream());
@@ -696,7 +698,8 @@ TYPED_TEST(StringToFloatTests, TrickyValues)
 
 TYPED_TEST(StringToFloatTests, Empty)
 {
-  auto empty = std::make_unique<column>(data_type{type_id::STRING}, 0, rmm::device_buffer{});
+  auto empty = std::make_unique<column>(data_type{type_id::STRING}, 0, rmm::device_buffer{},
+          rmm::device_buffer{}, 0);
 
   auto const result = spark_rapids_jni::string_to_float(data_type{type_to_id<TypeParam>()},
                                                         strings_column_view{empty->view()},

diff --git a/thirdparty/cudf b/thirdparty/cudf
+7 −3		cpp/benchmarks/common/generate_input.cu
+25 −57		cpp/include/cudf/column/column.hpp
+1 −1		cpp/src/column/column_factories.cpp
+2 −6		cpp/src/copying/contiguous_split.cu
+4 −2		cpp/src/datetime/timezone.cpp
+13 −4		cpp/src/interop/from_arrow.cu
+8 −3		cpp/src/io/json/json_column.cu
+2 −2		cpp/src/io/json/nested_json_gpu.cu
+2 −2		cpp/src/io/json/write_json.cu
+14 −2		cpp/src/io/parquet/page_enc.cu
+2 −2		cpp/src/io/utilities/column_buffer.cpp
+5 −2		cpp/src/lists/copying/copying.cu
+2 −1		cpp/src/partitioning/partitioning.cu
+2 −2		cpp/src/round/round.cu
+10 −4		cpp/src/text/subword/load_hash_file.cu
+4 −3		cpp/tests/bitmask/bitmask_tests.cpp
+55 −31		cpp/tests/column/column_test.cpp
+11 −6		cpp/tests/copying/concatenate_tests.cpp
+4 −5		cpp/tests/datetime/datetime_ops_test.cpp
+37 −0		cpp/tests/io/parquet_test.cpp
+2 −2		cpp/tests/structs/structs_column_tests.cpp
+31 −10		cpp/tests/unary/cast_tests.cpp
+99 −0		java/src/main/java/ai/rapids/cudf/ChunkedPack.java
+9 −18		java/src/main/java/ai/rapids/cudf/ContiguousTable.java
+74 −0		java/src/main/java/ai/rapids/cudf/PackedColumnMetadata.java
+100 −90		java/src/main/java/ai/rapids/cudf/Table.java
+57 −19		java/src/main/java/ai/rapids/cudf/TableWriter.java
+2 −0		java/src/main/native/CMakeLists.txt
+75 −0		java/src/main/native/src/ChunkedPackJni.cpp
+0 −20		java/src/main/native/src/ContiguousTableJni.cpp
+2 −2		java/src/main/native/src/CudfJni.cpp
+41 −0		java/src/main/native/src/PackedColumnMetadataJni.cpp
+90 −17		java/src/main/native/src/TableJni.cpp
+7 −6		java/src/main/native/src/row_conversion.cu
+113 −0		java/src/test/java/ai/rapids/cudf/TableTest.java
+6 −3		python/cudf/cudf/_lib/join.pyx
+14 −28		python/cudf/cudf/tests/test_dataframe_copy.py