This is an automated email from the ASF dual-hosted git repository.
felipecrv pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/arrow.git
The following commit(s) were added to refs/heads/main by this push:
new fb7e4682ca GH-41353: [C++] Define bit_width and byte_width of
ExtensionType in terms of the storage type (#41354)
fb7e4682ca is described below
commit fb7e4682ca5c19ec49997842650f80a453e0c1d1
Author: Felipe Oliveira Carvalho <[email protected]>
AuthorDate: Tue Apr 23 14:01:51 2024 -0300
GH-41353: [C++] Define bit_width and byte_width of ExtensionType in terms
of the storage type (#41354)
### Rationale for this change
Users and other classes within Arrow itself (e.g. array builders) expect
extension types to behave like their underlying storage type.
As it is now, `ExtensionType::bit_width()` is the default
`DataType::bit_width()` implementation which returns `-1`. It should return the
storage type's bit-width.
### What changes are included in this PR?
Definition of `ExtensionType::bit_width/byte_width` functions.
### Are these changes tested?
Tests added and confirmed to fail prior to these changes.
### Are there any user-facing changes?
`ExtensionType` now define `bit_width` and `byte_width` according to their
storage type.
* GitHub Issue: #41353
Authored-by: Felipe Oliveira Carvalho <[email protected]>
Signed-off-by: Felipe Oliveira Carvalho <[email protected]>
---
cpp/src/arrow/extension_type.h | 3 +++
cpp/src/arrow/extension_type_test.cc | 5 +++++
2 files changed, 8 insertions(+)
diff --git a/cpp/src/arrow/extension_type.h b/cpp/src/arrow/extension_type.h
index 0fd7216f18..b3f085198b 100644
--- a/cpp/src/arrow/extension_type.h
+++ b/cpp/src/arrow/extension_type.h
@@ -54,6 +54,9 @@ class ARROW_EXPORT ExtensionType : public DataType {
std::string name() const override { return "extension"; }
+ int32_t byte_width() const override { return storage_type_->byte_width(); }
+ int bit_width() const override { return storage_type_->bit_width(); }
+
/// \brief Unique name of extension type used to identify type for
/// serialization
/// \return the string name of the extension
diff --git a/cpp/src/arrow/extension_type_test.cc
b/cpp/src/arrow/extension_type_test.cc
index 31222d7480..f104c984a6 100644
--- a/cpp/src/arrow/extension_type_test.cc
+++ b/cpp/src/arrow/extension_type_test.cc
@@ -196,6 +196,8 @@ TEST_F(TestExtensionType, ExtensionTypeTest) {
auto type = uuid();
ASSERT_EQ(type->id(), Type::EXTENSION);
+ ASSERT_EQ(type->bit_width(), 128);
+ ASSERT_EQ(type->byte_width(), 16);
const auto& ext_type = static_cast<const ExtensionType&>(*type);
std::string serialized = ext_type.Serialize();
@@ -204,6 +206,9 @@ TEST_F(TestExtensionType, ExtensionTypeTest) {
ext_type.Deserialize(fixed_size_binary(16),
serialized));
ASSERT_TRUE(deserialized->Equals(*type));
ASSERT_FALSE(deserialized->Equals(*fixed_size_binary(16)));
+ ASSERT_EQ(deserialized->id(), Type::EXTENSION);
+ ASSERT_EQ(deserialized->bit_width(), 128);
+ ASSERT_EQ(deserialized->byte_width(), 16);
}
auto RoundtripBatch = [](const std::shared_ptr<RecordBatch>& batch,
