IcebergData fixes

sdks/julia/LogicalQueryProtocol.jl/src/properties.jl

@@ -149,12 +149,21 @@ function global_ids(data::Data)
             end
         end
         return ids
+    elseif dt.name == :iceberg_data
+        iceberg_data = dt[]::IcebergData
+        ids = LQPRelationId[]
+        for column in iceberg_data.columns
+            if !isnothing(column.target_id)
+                push!(ids, persistent_id(column.target_id))
+            end
+        end
+        return ids
     else
         @assert _is_valid_data(data)
     end
 end
 
 function _is_valid_data(data::Data)
     dt = data.data_type
-    return !isnothing(dt) && dt.name in [:edb, :betree_relation, :csv_data]
+    return !isnothing(dt) && dt.name in [:edb, :betree_relation, :csv_data, :iceberg_data]
 end

sdks/julia/LogicalQueryProtocol.jl/test/jet_test.jl

@@ -17,7 +17,8 @@
     # At runtime, the preceding `_has_proto_field` guard ensures correctness.
     function is_oneof_field_false_positive(report)
         msg = sprint(show, report)
-        return occursin("has no field args", msg) || occursin("has no field scale", msg)
+        return occursin("has no field args", msg) || occursin("has no field scale", msg) ||
+               occursin("has no field data_columns", msg)
     end
 
     # Filter BoundsError false positives from array internals in generated code.