Skip to content
Closed
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions cpp/src/arrow/CMakeLists.txt
Original file line number Diff line number Diff line change
Expand Up @@ -53,6 +53,7 @@ ADD_ARROW_TEST(array-list-test)
ADD_ARROW_TEST(array-primitive-test)
ADD_ARROW_TEST(array-string-test)
ADD_ARROW_TEST(array-struct-test)
ADD_ARROW_TEST(array-union-test)
ADD_ARROW_TEST(buffer-test)
ADD_ARROW_TEST(column-test)
ADD_ARROW_TEST(memory_pool-test)
Expand Down
62 changes: 42 additions & 20 deletions cpp/src/arrow/array-dictionary-test.cc
Original file line number Diff line number Diff line change
Expand Up @@ -34,7 +34,7 @@ namespace arrow {
TEST(TestDictionary, Basics) {
std::vector<int32_t> values = {100, 1000, 10000, 100000};
std::shared_ptr<Array> dict;
ArrayFromVector<Int32Type, int32_t>(int32(), values, &dict);
ArrayFromVector<Int32Type, int32_t>(values, &dict);

std::shared_ptr<DictionaryType> type1 =
std::dynamic_pointer_cast<DictionaryType>(dictionary(int16(), dict));
Expand All @@ -54,66 +54,88 @@ TEST(TestDictionary, Equals) {

std::shared_ptr<Array> dict;
std::vector<std::string> dict_values = {"foo", "bar", "baz"};
ArrayFromVector<StringType, std::string>(utf8(), dict_values, &dict);
ArrayFromVector<StringType, std::string>(dict_values, &dict);
std::shared_ptr<DataType> dict_type = dictionary(int16(), dict);

std::shared_ptr<Array> dict2;
std::vector<std::string> dict2_values = {"foo", "bar", "baz", "qux"};
ArrayFromVector<StringType, std::string>(utf8(), dict2_values, &dict2);
ArrayFromVector<StringType, std::string>(dict2_values, &dict2);
std::shared_ptr<DataType> dict2_type = dictionary(int16(), dict2);

std::shared_ptr<Array> indices;
std::vector<int16_t> indices_values = {1, 2, -1, 0, 2, 0};
ArrayFromVector<Int16Type, int16_t>(int16(), is_valid, indices_values, &indices);
ArrayFromVector<Int16Type, int16_t>(is_valid, indices_values, &indices);

std::shared_ptr<Array> indices2;
std::vector<int16_t> indices2_values = {1, 2, 0, 0, 2, 0};
ArrayFromVector<Int16Type, int16_t>(int16(), is_valid, indices2_values, &indices2);
ArrayFromVector<Int16Type, int16_t>(is_valid, indices2_values, &indices2);

std::shared_ptr<Array> indices3;
std::vector<int16_t> indices3_values = {1, 1, 0, 0, 2, 0};
ArrayFromVector<Int16Type, int16_t>(int16(), is_valid, indices3_values, &indices3);
ArrayFromVector<Int16Type, int16_t>(is_valid, indices3_values, &indices3);

auto arr = std::make_shared<DictionaryArray>(dict_type, indices);
auto arr2 = std::make_shared<DictionaryArray>(dict_type, indices2);
auto arr3 = std::make_shared<DictionaryArray>(dict2_type, indices);
auto arr4 = std::make_shared<DictionaryArray>(dict_type, indices3);
auto array = std::make_shared<DictionaryArray>(dict_type, indices);
auto array2 = std::make_shared<DictionaryArray>(dict_type, indices2);
auto array3 = std::make_shared<DictionaryArray>(dict2_type, indices);
auto array4 = std::make_shared<DictionaryArray>(dict_type, indices3);

ASSERT_TRUE(arr->Equals(arr));
ASSERT_TRUE(array->Equals(array));

// Equal, because the unequal index is masked by null
ASSERT_TRUE(arr->Equals(arr2));
ASSERT_TRUE(array->Equals(array2));

// Unequal dictionaries
ASSERT_FALSE(arr->Equals(arr3));
ASSERT_FALSE(array->Equals(array3));

// Unequal indices
ASSERT_FALSE(arr->Equals(arr4));
ASSERT_FALSE(array->Equals(array4));

// RangeEquals
ASSERT_TRUE(arr->RangeEquals(3, 6, 3, arr4));
ASSERT_FALSE(arr->RangeEquals(1, 3, 1, arr4));
ASSERT_TRUE(array->RangeEquals(3, 6, 3, array4));
ASSERT_FALSE(array->RangeEquals(1, 3, 1, array4));

// ARROW-33 Test slices
const int size = array->length();

std::shared_ptr<Array> slice, slice2;
slice = array->Array::Slice(2);
slice2 = array->Array::Slice(2);
ASSERT_EQ(size - 2, slice->length());

ASSERT_TRUE(slice->Equals(slice2));
ASSERT_TRUE(array->RangeEquals(2, array->length(), 0, slice));

// Chained slices
slice2 = array->Array::Slice(1)->Array::Slice(1);
ASSERT_TRUE(slice->Equals(slice2));

slice = array->Slice(1, 3);
slice2 = array->Slice(1, 3);
ASSERT_EQ(3, slice->length());

ASSERT_TRUE(slice->Equals(slice2));
ASSERT_TRUE(array->RangeEquals(1, 4, 0, slice));
}

TEST(TestDictionary, Validate) {
std::vector<bool> is_valid = {true, true, false, true, true, true};

std::shared_ptr<Array> dict;
std::vector<std::string> dict_values = {"foo", "bar", "baz"};
ArrayFromVector<StringType, std::string>(utf8(), dict_values, &dict);
ArrayFromVector<StringType, std::string>(dict_values, &dict);
std::shared_ptr<DataType> dict_type = dictionary(int16(), dict);

std::shared_ptr<Array> indices;
std::vector<uint8_t> indices_values = {1, 2, 0, 0, 2, 0};
ArrayFromVector<UInt8Type, uint8_t>(uint8(), is_valid, indices_values, &indices);
ArrayFromVector<UInt8Type, uint8_t>(is_valid, indices_values, &indices);

std::shared_ptr<Array> indices2;
std::vector<float> indices2_values = {1., 2., 0., 0., 2., 0.};
ArrayFromVector<FloatType, float>(float32(), is_valid, indices2_values, &indices2);
ArrayFromVector<FloatType, float>(is_valid, indices2_values, &indices2);

std::shared_ptr<Array> indices3;
std::vector<int64_t> indices3_values = {1, 2, 0, 0, 2, 0};
ArrayFromVector<Int64Type, int64_t>(int64(), is_valid, indices3_values, &indices3);
ArrayFromVector<Int64Type, int64_t>(is_valid, indices3_values, &indices3);

std::shared_ptr<Array> arr = std::make_shared<DictionaryArray>(dict_type, indices);
std::shared_ptr<Array> arr2 = std::make_shared<DictionaryArray>(dict_type, indices2);
Expand Down
36 changes: 28 additions & 8 deletions cpp/src/arrow/array-list-test.cc
Original file line number Diff line number Diff line change
Expand Up @@ -90,9 +90,9 @@ TEST_F(TestListBuilder, Equality) {
Int32Builder* vb = static_cast<Int32Builder*>(builder_->value_builder().get());

std::shared_ptr<Array> array, equal_array, unequal_array;
vector<int32_t> equal_offsets = {0, 1, 2, 5};
vector<int32_t> equal_values = {1, 2, 3, 4, 5, 2, 2, 2};
vector<int32_t> unequal_offsets = {0, 1, 4};
vector<int32_t> equal_offsets = {0, 1, 2, 5, 6, 7, 8, 10};
vector<int32_t> equal_values = {1, 2, 3, 4, 5, 2, 2, 2, 5, 6};
vector<int32_t> unequal_offsets = {0, 1, 4, 7};
vector<int32_t> unequal_values = {1, 2, 2, 2, 3, 4, 5};

// setup two equal arrays
Expand Down Expand Up @@ -122,7 +122,27 @@ TEST_F(TestListBuilder, Equality) {
EXPECT_FALSE(array->RangeEquals(0, 2, 0, unequal_array));
EXPECT_FALSE(array->RangeEquals(1, 2, 1, unequal_array));
EXPECT_TRUE(array->RangeEquals(2, 3, 2, unequal_array));
EXPECT_TRUE(array->RangeEquals(3, 4, 1, unequal_array));

// Check with slices, ARROW-33
std::shared_ptr<Array> slice, slice2;

slice = array->Slice(2);
slice2 = array->Slice(2);
ASSERT_EQ(array->length() - 2, slice->length());

ASSERT_TRUE(slice->Equals(slice2));
ASSERT_TRUE(array->RangeEquals(2, slice->length(), 0, slice));

// Chained slices
slice2 = array->Slice(1)->Slice(1);
ASSERT_TRUE(slice->Equals(slice2));

slice = array->Slice(1, 4);
slice2 = array->Slice(1, 4);
ASSERT_EQ(4, slice->length());

ASSERT_TRUE(slice->Equals(slice2));
ASSERT_TRUE(array->RangeEquals(1, 5, 0, slice));
}

TEST_F(TestListBuilder, TestResize) {}
Expand All @@ -137,9 +157,9 @@ TEST_F(TestListBuilder, TestAppendNull) {
ASSERT_TRUE(result_->IsNull(0));
ASSERT_TRUE(result_->IsNull(1));

ASSERT_EQ(0, result_->raw_offsets()[0]);
ASSERT_EQ(0, result_->offset(1));
ASSERT_EQ(0, result_->offset(2));
ASSERT_EQ(0, result_->raw_value_offsets()[0]);
ASSERT_EQ(0, result_->value_offset(1));
ASSERT_EQ(0, result_->value_offset(2));

Int32Array* values = static_cast<Int32Array*>(result_->values().get());
ASSERT_EQ(0, values->length());
Expand All @@ -154,7 +174,7 @@ void ValidateBasicListArray(const ListArray* result, const vector<int32_t>& valu
ASSERT_EQ(3, result->length());
vector<int32_t> ex_offsets = {0, 3, 3, 7};
for (size_t i = 0; i < ex_offsets.size(); ++i) {
ASSERT_EQ(ex_offsets[i], result->offset(i));
ASSERT_EQ(ex_offsets[i], result->value_offset(i));
}

for (int i = 0; i < result->length(); ++i) {
Expand Down
78 changes: 73 additions & 5 deletions cpp/src/arrow/array-primitive-test.cc
Original file line number Diff line number Diff line change
Expand Up @@ -121,7 +121,7 @@ class TestPrimitiveBuilder : public TestBuilder {
}

auto expected =
std::make_shared<ArrayType>(size, ex_data, ex_null_count, ex_null_bitmap);
std::make_shared<ArrayType>(size, ex_data, ex_null_bitmap, ex_null_count);

std::shared_ptr<Array> out;
ASSERT_OK(builder->Finish(&out));
Expand Down Expand Up @@ -217,7 +217,7 @@ void TestPrimitiveBuilder<PBoolean>::Check(
}

auto expected =
std::make_shared<BooleanArray>(size, ex_data, ex_null_count, ex_null_bitmap);
std::make_shared<BooleanArray>(size, ex_data, ex_null_bitmap, ex_null_count);

std::shared_ptr<Array> out;
ASSERT_OK(builder->Finish(&out));
Expand All @@ -235,15 +235,14 @@ void TestPrimitiveBuilder<PBoolean>::Check(

for (int i = 0; i < result->length(); ++i) {
if (nullable) { ASSERT_EQ(valid_bytes_[i] == 0, result->IsNull(i)) << i; }
bool actual = BitUtil::GetBit(result->raw_data(), i);
bool actual = BitUtil::GetBit(result->data()->data(), i);
ASSERT_EQ(static_cast<bool>(draws_[i]), actual) << i;
}
ASSERT_TRUE(result->Equals(*expected));
}

typedef ::testing::Types<PBoolean, PUInt8, PUInt16, PUInt32, PUInt64, PInt8, PInt16,
PInt32, PInt64, PFloat, PDouble>
Primitives;
PInt32, PInt64, PFloat, PDouble> Primitives;

TYPED_TEST_CASE(TestPrimitiveBuilder, Primitives);

Expand Down Expand Up @@ -347,6 +346,39 @@ TYPED_TEST(TestPrimitiveBuilder, Equality) {
array->RangeEquals(first_valid_idx + 1, size, first_valid_idx + 1, unequal_array));
}

TYPED_TEST(TestPrimitiveBuilder, SliceEquality) {
DECL_T();

const int size = 1000;
this->RandomData(size);
vector<T>& draws = this->draws_;
vector<uint8_t>& valid_bytes = this->valid_bytes_;
auto builder = this->builder_.get();

std::shared_ptr<Array> array;
ASSERT_OK(MakeArray(valid_bytes, draws, size, builder, &array));

std::shared_ptr<Array> slice, slice2;

slice = array->Slice(5);
slice2 = array->Slice(5);
ASSERT_EQ(size - 5, slice->length());

ASSERT_TRUE(slice->Equals(slice2));
ASSERT_TRUE(array->RangeEquals(5, array->length(), 0, slice));

// Chained slices
slice2 = array->Slice(2)->Slice(3);
ASSERT_TRUE(slice->Equals(slice2));

slice = array->Slice(5, 10);
slice2 = array->Slice(5, 10);
ASSERT_EQ(10, slice->length());

ASSERT_TRUE(slice->Equals(slice2));
ASSERT_TRUE(array->RangeEquals(5, 15, 0, slice));
}

TYPED_TEST(TestPrimitiveBuilder, TestAppendScalar) {
DECL_T();

Expand Down Expand Up @@ -473,4 +505,40 @@ TYPED_TEST(TestPrimitiveBuilder, TestReserve) {
ASSERT_EQ(BitUtil::NextPower2(kMinBuilderCapacity + 100), this->builder_->capacity());
}

template <typename TYPE>
void CheckSliceApproxEquals() {
using T = typename TYPE::c_type;

const int kSize = 50;
std::vector<T> draws1;
std::vector<T> draws2;

const uint32_t kSeed = 0;
test::random_real<T>(kSize, kSeed, 0, 100, &draws1);
test::random_real<T>(kSize, kSeed + 1, 0, 100, &draws2);

// Make the draws equal in the sliced segment, but unequal elsewhere (to
// catch not using the slice offset)
for (int i = 10; i < 30; ++i) {
draws2[i] = draws1[i];
}

std::vector<bool> is_valid;
test::random_is_valid(kSize, 0.1, &is_valid);

std::shared_ptr<Array> array1, array2;
ArrayFromVector<TYPE, T>(is_valid, draws1, &array1);
ArrayFromVector<TYPE, T>(is_valid, draws2, &array2);

std::shared_ptr<Array> slice1 = array1->Slice(10, 20);
std::shared_ptr<Array> slice2 = array2->Slice(10, 20);

ASSERT_TRUE(slice1->ApproxEquals(slice2));
}

TEST(TestPrimitiveAdHoc, FloatingSliceApproxEquals) {
CheckSliceApproxEquals<FloatType>();
CheckSliceApproxEquals<DoubleType>();
}

} // namespace arrow
Loading