From 63c82cb76cf9d39469353f14a14a8fe1d98c87df Mon Sep 17 00:00:00 2001
From: Christian Bush <chbush@linkedin.com>
Date: Fri, 20 Feb 2026 14:16:05 -0800
Subject: [PATCH] Task #1: Add OrcSchemaManifest and OrcSchemaField structures

- Added OrcSchemaField struct to map Arrow fields to ORC column indices
- Added OrcSchemaManifest struct for schema mapping infrastructure
- Includes GetColumnField() and GetParent() helper methods
- Added stub Make() implementation (full logic in Task #2)
- Mirrors Parquet SchemaManifest design adapted for ORC type system

Verified: Code structure matches Parquet pattern

Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
---
 cpp/src/arrow/dataset/file_orc.cc | 10 +++++
 cpp/src/arrow/dataset/file_orc.h  | 69 +++++++++++++++++++++++++++++++
 2 files changed, 79 insertions(+)

diff --git a/cpp/src/arrow/dataset/file_orc.cc b/cpp/src/arrow/dataset/file_orc.cc
index 1393df57f9d7..8cb44cd5c0be 100644
--- a/cpp/src/arrow/dataset/file_orc.cc
+++ b/cpp/src/arrow/dataset/file_orc.cc
@@ -35,6 +35,16 @@ using internal::checked_pointer_cast;
 
 namespace dataset {
 
+// OrcSchemaManifest implementation
+Status OrcSchemaManifest::Make(const std::shared_ptr<Schema>& schema,
+                                const void* orc_type, OrcSchemaManifest* manifest) {
+  // TODO(Task #2): Implement BuildOrcSchemaManifest logic
+  // This is a placeholder for Task #1 - actual implementation in Task #2
+  manifest->origin_schema = schema;
+  return Status::NotImplemented(
+      "OrcSchemaManifest::Make will be implemented in Task #2");
+}
+
 namespace {
 
 Result<std::unique_ptr<arrow::adapters::orc::ORCFileReader>> OpenORCReader(
diff --git a/cpp/src/arrow/dataset/file_orc.h b/cpp/src/arrow/dataset/file_orc.h
index 5bfefd1e02b5..3c36dcd5d8d9 100644
--- a/cpp/src/arrow/dataset/file_orc.h
+++ b/cpp/src/arrow/dataset/file_orc.h
@@ -21,12 +21,16 @@
 
 #include <memory>
 #include <string>
+#include <unordered_map>
+#include <vector>
 
 #include "arrow/dataset/file_base.h"
 #include "arrow/dataset/type_fwd.h"
 #include "arrow/dataset/visibility.h"
 #include "arrow/io/type_fwd.h"
 #include "arrow/result.h"
+#include "arrow/status.h"
+#include "arrow/type_fwd.h"
 
 namespace arrow {
 namespace dataset {
@@ -35,6 +39,71 @@ namespace dataset {
 ///
 /// @{
 
+/// \brief Bridge between an arrow::Field and ORC column indices.
+///
+/// Similar to Parquet's SchemaField, this structure maps Arrow schema fields
+/// to ORC physical column indices. ORC uses a depth-first pre-order traversal
+/// where column 0 is the root struct, and subsequent indices are for child columns.
+struct ARROW_DS_EXPORT OrcSchemaField {
+  /// The Arrow field corresponding to this ORC column
+  std::shared_ptr<Field> field;
+
+  /// Child fields (for nested types like structs, lists, maps)
+  std::vector<OrcSchemaField> children;
+
+  /// ORC column index (only set for leaf nodes that have statistics)
+  /// For ORC, column 0 is the root struct, columns 1+ are the actual data columns
+  int column_index = -1;
+
+  /// Check if this is a leaf node (has column statistics)
+  bool is_leaf() const { return column_index != -1; }
+};
+
+/// \brief Bridge between an ORC file schema and an Arrow Schema.
+///
+/// Maps Arrow schema fields to ORC physical column indices for statistics lookup.
+/// Similar to Parquet's SchemaManifest but adapted for ORC's type system.
+struct ARROW_DS_EXPORT OrcSchemaManifest {
+  /// Create a schema manifest from ORC type information
+  /// \param schema The Arrow schema
+  /// \param orc_type Pointer to orc::Type from the ORC reader (as void* to avoid ORC header dependency)
+  /// \param manifest Output manifest to populate
+  static Status Make(const std::shared_ptr<Schema>& schema, const void* orc_type,
+                     OrcSchemaManifest* manifest);
+
+  /// The Arrow schema
+  std::shared_ptr<Schema> origin_schema;
+
+  /// Top-level schema fields
+  std::vector<OrcSchemaField> schema_fields;
+
+  /// Map from ORC column index to schema field (for fast lookup)
+  std::unordered_map<int, const OrcSchemaField*> column_index_to_field;
+
+  /// Map from child field to parent field (for traversal)
+  std::unordered_map<const OrcSchemaField*, const OrcSchemaField*> child_to_parent;
+
+  /// Get the schema field for a given ORC column index
+  Status GetColumnField(int column_index, const OrcSchemaField** out) const {
+    auto it = column_index_to_field.find(column_index);
+    if (it == column_index_to_field.end()) {
+      return Status::KeyError("Column index ", column_index,
+                             " not found in ORC schema manifest");
+    }
+    *out = it->second;
+    return Status::OK();
+  }
+
+  /// Get the parent field of a given field
+  const OrcSchemaField* GetParent(const OrcSchemaField* field) const {
+    auto it = child_to_parent.find(field);
+    if (it == child_to_parent.end()) {
+      return nullptr;
+    }
+    return it->second;
+  }
+};
+
 constexpr char kOrcTypeName[] = "orc";
 
 /// \brief A FileFormat implementation that reads from and writes to ORC files