OGRLayer::GetArrowStream(): when DATETIME_AS_STRING=YES, expose "GDAL…

…:OGR:type":"DateTime" metadata in the ArrowSchema of DateTime fields
rouault · Nov 7, 2024 · d881e74 · d881e74
1 parent ad41bb8
commit d881e74
Show file tree

Hide file tree

Showing 3 changed files with 55 additions and 1 deletion.
diff --git a/autotest/cpp/CMakeLists.txt b/autotest/cpp/CMakeLists.txt
@@ -108,7 +108,7 @@ gdal_standard_includes(gdal_unit_test)
 target_compile_options(gdal_unit_test PRIVATE ${GDAL_CXX_WARNING_FLAGS})
 target_compile_definitions(gdal_unit_test PRIVATE -DGDAL_TEST_ROOT_DIR="${GDAL_ROOT_TEST_DIR}")
 target_include_directories(
-  gdal_unit_test PRIVATE $<TARGET_PROPERTY:appslib,SOURCE_DIR> $<TARGET_PROPERTY:gdal_vrt,SOURCE_DIR>)
+  gdal_unit_test PRIVATE $<TARGET_PROPERTY:appslib,SOURCE_DIR> $<TARGET_PROPERTY:gdal_vrt,SOURCE_DIR> $<TARGET_PROPERTY:ogrsf_generic,SOURCE_DIR>)
 if (GDAL_USE_SQLITE3)
   target_compile_definitions(gdal_unit_test PRIVATE -DHAVE_SQLITE3)
   target_include_directories(

diff --git a/autotest/cpp/test_ogr.cpp b/autotest/cpp/test_ogr.cpp
@@ -16,6 +16,7 @@
 #include "ogrsf_frmts.h"
 #include "../../ogr/ogrsf_frmts/osm/gpb.h"
 #include "ogr_recordbatch.h"
+#include "ogrlayerarrow.h"
 
 #include <string>
 #include <algorithm>
@@ -4444,4 +4445,43 @@ TEST_F(test_ogr, OGRFeature_SetGeomField)
     poFeatureDefn->Release();
 }
 
+TEST_F(test_ogr, GetArrowStream_DateTime_As_String)
+{
+    auto poDS = std::unique_ptr<GDALDataset>(
+        GetGDALDriverManager()->GetDriverByName("Memory")->Create(
+            "", 0, 0, 0, GDT_Unknown, nullptr));
+    auto poLayer = poDS->CreateLayer("test", nullptr, wkbNone);
+    OGRFieldDefn oFieldDefn("dt", OFTDateTime);
+    poLayer->CreateField(&oFieldDefn);
+    struct ArrowArrayStream stream;
+    CPLStringList aosOptions;
+    aosOptions.SetNameValue("INCLUDE_FID", "NO");
+    aosOptions.SetNameValue("DATETIME_AS_STRING", "YES");
+    ASSERT_TRUE(poLayer->GetArrowStream(&stream, aosOptions.List()));
+    struct ArrowSchema schema;
+    memset(&schema, 0, sizeof(schema));
+    EXPECT_EQ(stream.get_schema(&stream, &schema), 0);
+    EXPECT_TRUE(schema.n_children == 1 &&
+                strcmp(schema.children[0]->format, "u") == 0)
+        << schema.n_children;
+    if (schema.n_children == 1 && strcmp(schema.children[0]->format, "u") == 0)
+    {
+        EXPECT_TRUE(schema.children[0]->metadata != nullptr);
+        if (schema.children[0]->metadata)
+        {
+            auto oMapKeyValue =
+                OGRParseArrowMetadata(schema.children[0]->metadata);
+            EXPECT_EQ(oMapKeyValue.size(), 1);
+            if (oMapKeyValue.size() == 1)
+            {
+                EXPECT_STREQ(oMapKeyValue.begin()->first.c_str(),
+                             "GDAL:OGR:type");
+                EXPECT_STREQ(oMapKeyValue.begin()->second.c_str(), "DateTime");
+            }
+        }
+    }
+    schema.release(&schema);
+    stream.release(&stream);
+}
+
 }  // namespace
diff --git a/ogr/ogrsf_frmts/generic/ogrlayerarrow.cpp b/ogr/ogrsf_frmts/generic/ogrlayerarrow.cpp
@@ -32,6 +32,7 @@
 #include <utility>
 #include <set>
 
+constexpr const char *MD_GDAL_OGR_TYPE = "GDAL:OGR:type";
 constexpr const char *MD_GDAL_OGR_ALTERNATIVE_NAME =
     "GDAL:OGR:alternative_name";
 constexpr const char *MD_GDAL_OGR_COMMENT = "GDAL:OGR:comment";
@@ -579,6 +580,13 @@ int OGRLayer::GetArrowSchema(struct ArrowArrayStream *,
         }
 
         std::vector<std::pair<std::string, std::string>> oMetadata;
+
+        if (eType == OFTDateTime && bDateTimeAsString)
+        {
+            oMetadata.emplace_back(
+                std::pair(MD_GDAL_OGR_TYPE, OGR_GetFieldTypeName(eType)));
+        }
+
         const char *pszAlternativeName = poFieldDefn->GetAlternativeNameRef();
         if (pszAlternativeName && pszAlternativeName[0])
             oMetadata.emplace_back(
@@ -2472,6 +2480,9 @@ const char *OGRLayer::GetLastErrorArrowArrayStream(struct ArrowArrayStream *)
  * Starting with GDAL 3.8, the ArrowSchema::metadata field filled by the
  * get_schema() callback may be set with the potential following items:
  * <ul>
+ * <li>"GDAL:OGR:type": value of OGRFieldDefn::GetType(): (added in 3.11)
+ *      Only used for DateTime fields when the DATETIME_AS_STRING=YES option is
+ *      specified.</li>
  * <li>"GDAL:OGR:alternative_name": value of
  *     OGRFieldDefn::GetAlternativeNameRef()</li>
  * <li>"GDAL:OGR:comment": value of OGRFieldDefn::GetComment()</li>
@@ -2674,6 +2685,9 @@ bool OGRLayer::GetArrowStream(struct ArrowArrayStream *out_stream,
  * Starting with GDAL 3.8, the ArrowSchema::metadata field filled by the
  * get_schema() callback may be set with the potential following items:
  * <ul>
+ * <li>"GDAL:OGR:type": value of OGRFieldDefn::GetType(): (added in 3.11)
+ *      Only used for DateTime fields when the DATETIME_AS_STRING=YES option is
+ *      specified.</li>
  * <li>"GDAL:OGR:alternative_name": value of
  *     OGRFieldDefn::GetAlternativeNameRef()</li>
  * <li>"GDAL:OGR:comment": value of OGRFieldDefn::GetComment()</li>