[c++] Merges for #1559

apis/python/src/tiledbsoma/pytiledbsoma.cc

@@ -129,9 +129,9 @@ py::object to_table(SOMAArray& sr, std::shared_ptr<ArrayBuffers> array_buffers)
             arrays.append(to_array(column));
         }else{
             arrays.append(pa_dict_from_arrays(
-                to_array(column), 
-                get_enum(sr, name), 
-                py::none(), 
+                to_array(column),
+                get_enum(sr, name),
+                py::none(),
                 get_enum_is_ordered(sr, name)));
         }
     }
@@ -640,11 +640,11 @@ PYBIND11_MODULE(pytiledbsoma, m) {
         .def("nnz", &SOMAArray::nnz, py::call_guard<py::gil_scoped_release>())
 
         .def_property_readonly("shape", &SOMAArray::shape)
-        
+
         .def("get_enum", get_enum)
 
         .def("get_enum_is_ordered", get_enum_is_ordered)
-        
+
         .def("get_enum_label_on_attr", &SOMAArray::get_enum_label_on_attr);
 }
 }  // namespace tiledbsoma

apis/r/src/rinterface.cpp

@@ -148,7 +148,7 @@ Rcpp::List soma_array_reader(const std::string& uri,
         auto buf = sr_data->get()->at(names[i]);
 
         // this is pair of array and schema pointer
-        auto pp = tdbs::ArrowAdapter::to_arrow(buf);
+        auto pp = tdbs::ArrowAdapter::to_arrow(buf, true);
 
         memcpy((void*) chldschemaxp, pp.second.get(), sizeof(ArrowSchema));
         memcpy((void*) chldarrayxp, pp.first.get(), sizeof(ArrowArray));

apis/r/src/riterator.cpp

@@ -216,7 +216,7 @@ Rcpp::List sr_next(Rcpp::XPtr<tdbs::SOMAArray> sr) {
        auto buf = sr_data->get()->at(names[i]);
 
        // this is pair of array and schema pointer
-       auto pp = tdbs::ArrowAdapter::to_arrow(buf);
+       auto pp = tdbs::ArrowAdapter::to_arrow(buf, true);
 
        memcpy((void*) chldschemaxp, pp.second.get(), sizeof(ArrowSchema));
        memcpy((void*) chldarrayxp, pp.first.get(), sizeof(ArrowArray));

libtiledbsoma/src/soma/managed_query.cc

@@ -225,6 +225,9 @@ std::shared_ptr<ArrayBuffers> ManagedQuery::results() {
             auto enumeration = ArrayExperimental::get_enumeration(
                 *ctx_, *array_, enumname.value());
             auto enumvec = enumeration.as_vector<std::string>();
+            if (!buffers_->contains(attrname)) {
+                continue;
+            }
             auto colbuf = buffers_->at(attrname);
             colbuf->add_enumeration(enumvec);
             LOG_DEBUG(fmt::format(

libtiledbsoma/src/soma/soma_array.cc

@@ -571,4 +571,4 @@ void SOMAArray::validate(
     }
 }
 
-}  // namespace tiledbsoma
+}  // namespace tiledbsoma

libtiledbsoma/src/utils/arrow_adapter.h

@@ -86,7 +86,8 @@ class ArrowAdapter {
      *
      * @return auto [Arrow array, Arrow schema]
      */
-    static auto to_arrow(std::shared_ptr<ColumnBuffer> column) {
+    static auto to_arrow(
+        std::shared_ptr<ColumnBuffer> column, bool use_enum = false) {
         std::unique_ptr<ArrowSchema> schema = std::make_unique<ArrowSchema>();
         std::unique_ptr<ArrowArray> array = std::make_unique<ArrowArray>();
 
@@ -153,8 +154,10 @@ class ArrowAdapter {
             column->data_to_bitmap();
         }
 
-        // If we have an enumeration, fill a dictionary
-        if (column->has_enumeration()) {
+        // If we have an enumeration, fill a dictionary.
+        // The Python callpath handles this separately. The R callpath needs us
+        // to do this. TODO: uniformize this at the callsites.
+        if (column->has_enumeration() && use_enum) {
             auto enumvec = column->get_enumeration();
 
             ArrowSchema* dict_sch = new ArrowSchema;