4d158f9050
Add a new function in arrow adapter to get the raw orc reader which we can get more information from such offset or min/max value. And this will be used in #1046 This modify is inspired by Clickhouse
70 lines
2.2 KiB
Diff
70 lines
2.2 KiB
Diff
diff --git a/cpp/src/arrow/adapters/orc/adapter.cc b/cpp/src/arrow/adapters/orc/adapter.cc
|
|
index 03243e7..cbb3ed9 100644
|
|
--- a/cpp/src/arrow/adapters/orc/adapter.cc
|
|
+++ b/cpp/src/arrow/adapters/orc/adapter.cc
|
|
@@ -47,9 +47,6 @@
|
|
#include "arrow/util/visibility.h"
|
|
#include "orc/Exceptions.hh"
|
|
|
|
-// alias to not interfere with nested orc namespace
|
|
-namespace liborc = orc;
|
|
-
|
|
#define ORC_THROW_NOT_OK(s) \
|
|
do { \
|
|
Status _s = (s); \
|
|
@@ -198,6 +195,11 @@ class ORCFileReader::Impl {
|
|
return Init();
|
|
}
|
|
|
|
+ virtual liborc::Reader* GetRawORCReader() {
|
|
+ return reader_.get();
|
|
+ }
|
|
+
|
|
+
|
|
Status Init() {
|
|
int64_t nstripes = reader_->getNumberOfStripes();
|
|
stripes_.resize(nstripes);
|
|
@@ -504,6 +506,7 @@ class ORCFileReader::Impl {
|
|
return Status::OK();
|
|
}
|
|
|
|
+
|
|
Status NextStripeReader(int64_t batch_size, std::shared_ptr<RecordBatchReader>* out) {
|
|
return NextStripeReader(batch_size, {}, out);
|
|
}
|
|
@@ -531,6 +534,10 @@ Result<std::unique_ptr<ORCFileReader>> ORCFileReader::Open(
|
|
return std::move(result);
|
|
}
|
|
|
|
+liborc::Reader* ORCFileReader::GetRawORCReader() {
|
|
+ return impl_->GetRawORCReader();
|
|
+}
|
|
+
|
|
Result<std::shared_ptr<const KeyValueMetadata>> ORCFileReader::ReadMetadata() {
|
|
return impl_->ReadMetadata();
|
|
}
|
|
diff --git a/cpp/src/arrow/adapters/orc/adapter.h b/cpp/src/arrow/adapters/orc/adapter.h
|
|
index 223efa5..a0d112a 100644
|
|
--- a/cpp/src/arrow/adapters/orc/adapter.h
|
|
+++ b/cpp/src/arrow/adapters/orc/adapter.h
|
|
@@ -30,6 +30,10 @@
|
|
#include "arrow/type_fwd.h"
|
|
#include "arrow/util/macros.h"
|
|
#include "arrow/util/visibility.h"
|
|
+#include "orc/Reader.hh"
|
|
+
|
|
+// alias to not interfere with nested orc namespace
|
|
+namespace liborc = orc;
|
|
|
|
namespace arrow {
|
|
namespace adapters {
|
|
@@ -50,6 +54,9 @@ class ARROW_EXPORT ORCFileReader {
|
|
ARROW_DEPRECATED("Deprecated in 6.0.0. Use Result-returning overload instead.")
|
|
static Status Open(const std::shared_ptr<io::RandomAccessFile>& file, MemoryPool* pool,
|
|
std::unique_ptr<ORCFileReader>* reader);
|
|
+
|
|
+ /// \brief Get ORC reader from inside.
|
|
+ liborc::Reader* GetRawORCReader();
|
|
|
|
/// \brief Creates a new ORC reader
|
|
///
|