GH=36099: [C++] Add Utf8View and BinaryView to the c data bridge

bkietz · Oct 24, 2023 · b0bcb16 · b0bcb16
1 parent 6f2e851
commit b0bcb16
Show file tree

Hide file tree

Showing 3 changed files with 32 additions and 2 deletions.
diff --git a/cpp/src/arrow/c/bridge.cc b/cpp/src/arrow/c/bridge.cc
@@ -20,6 +20,7 @@
 #include <algorithm>
 #include <cerrno>
 #include <cstring>
+#include <memory>
 #include <string>
 #include <string_view>
 #include <utility>
@@ -260,7 +261,7 @@ struct SchemaExporter {
       // Dictionary type: parent struct describes index type,
       // child dictionary struct describes value type.
       RETURN_NOT_OK(VisitTypeInline(*dict_type.index_type(), this));
-      dict_exporter_.reset(new SchemaExporter());
+      dict_exporter_ = std::make_unique<SchemaExporter>();
       RETURN_NOT_OK(dict_exporter_->ExportType(*dict_type.value_type()));
     } else {
       RETURN_NOT_OK(VisitTypeInline(type, this));
@@ -357,10 +358,14 @@ struct SchemaExporter {
 
   Status Visit(const LargeBinaryType& type) { return SetFormat("Z"); }
 
+  Status Visit(const BinaryViewType& type) { return SetFormat("vz"); }
+
   Status Visit(const StringType& type) { return SetFormat("u"); }
 
   Status Visit(const LargeStringType& type) { return SetFormat("U"); }
 
+  Status Visit(const StringViewType& type) { return SetFormat("vu"); }
+
   Status Visit(const Date32Type& type) { return SetFormat("tdD"); }
 
   Status Visit(const Date64Type& type) { return SetFormat("tdm"); }
@@ -574,7 +579,7 @@ struct ArrayExporter {
 
     // Export dictionary
     if (data->dictionary != nullptr) {
-      dict_exporter_.reset(new ArrayExporter());
+      dict_exporter_ = std::make_unique<ArrayExporter>();
       RETURN_NOT_OK(dict_exporter_->Export(data->dictionary));
     }
 
@@ -1064,6 +1069,8 @@ struct SchemaImporter {
         return ProcessPrimitive(binary());
       case 'Z':
         return ProcessPrimitive(large_binary());
+      case 'v':
+        return ProcessBinaryView();
       case 'w':
         return ProcessFixedSizeBinary();
       case 'd':
@@ -1076,6 +1083,17 @@ struct SchemaImporter {
     return f_parser_.Invalid();
   }
 
+  Status ProcessBinaryView() {
+    RETURN_NOT_OK(f_parser_.CheckHasNext());
+    switch (f_parser_.Next()) {
+      case 'z':
+        return ProcessPrimitive(binary_view());
+      case 'u':
+        return ProcessPrimitive(utf8_view());
+    }
+    return f_parser_.Invalid();
+  }
+
   Status ProcessTemporal() {
     RETURN_NOT_OK(f_parser_.CheckHasNext());
     switch (f_parser_.Next()) {

diff --git a/cpp/src/arrow/c/bridge_test.cc b/cpp/src/arrow/c/bridge_test.cc
@@ -353,6 +353,8 @@ TEST_F(TestSchemaExport, Primitive) {
   TestPrimitive(large_binary(), "Z");
   TestPrimitive(utf8(), "u");
   TestPrimitive(large_utf8(), "U");
+  TestPrimitive(binary_view(), "vz");
+  TestPrimitive(utf8_view(), "vu");
 
   TestPrimitive(decimal(16, 4), "d:16,4");
   TestPrimitive(decimal256(16, 4), "d:16,4,256");
@@ -874,6 +876,8 @@ TEST_F(TestArrayExport, Primitive) {
   TestPrimitive(large_binary(), R"(["foo", "bar", null])");
   TestPrimitive(utf8(), R"(["foo", "bar", null])");
   TestPrimitive(large_utf8(), R"(["foo", "bar", null])");
+  TestPrimitive(binary_view(), R"(["foo", "bar", null])");
+  TestPrimitive(utf8_view(), R"(["foo", "bar", null])");
 
   TestPrimitive(decimal(16, 4), R"(["1234.5670", null])");
   TestPrimitive(decimal256(16, 4), R"(["1234.5670", null])");
@@ -1891,6 +1895,10 @@ TEST_F(TestSchemaImport, String) {
   CheckImport(large_utf8());
   FillPrimitive("Z");
   CheckImport(large_binary());
+  FillPrimitive("vu");
+  CheckImport(utf8_view());
+  FillPrimitive("vz");
+  CheckImport(binary_view());
 
   FillPrimitive("w:3");
   CheckImport(fixed_size_binary(3));

diff --git a/docs/source/format/CDataInterface.rst b/docs/source/format/CDataInterface.rst
@@ -140,10 +140,14 @@ strings:
 +-----------------+---------------------------------------------------+------------+
 | ``Z``           | large binary                                      |            |
 +-----------------+---------------------------------------------------+------------+
+| ``vz``          | binary view                                       |            |
++-----------------+---------------------------------------------------+------------+
 | ``u``           | utf-8 string                                      |            |
 +-----------------+---------------------------------------------------+------------+
 | ``U``           | large utf-8 string                                |            |
 +-----------------+---------------------------------------------------+------------+
+| ``vu``          | utf-8 view                                        |            |
++-----------------+---------------------------------------------------+------------+
 | ``d:19,10``     | decimal128 [precision 19, scale 10]               |            |
 +-----------------+---------------------------------------------------+------------+
 | ``d:19,10,NNN`` | decimal bitwidth = NNN [precision 19, scale 10]   |            |