Back
Coordinates Decomposer
Overview
The Coordinates decomposer specializes in extracting spatial coordinate components
from vector types. Unlike Components, it uses an underscore separator and allows
customizable coordinate names (defaulting to x, y, z).
Include
#include <dataframe/attributes/Coordinates.h>Class Definition
Coordinates API
namespace df::attributes {
class Coordinates : public GenDecomposer<Coordinates> {
public:
// Construct with custom coordinate names (default: {"x", "y", "z"})
explicit Coordinates(
const std::vector<std::string>& coordNames = {"x", "y", "z"});
Strings names(
const Dataframe& dataframe,
DecompDimension targetDim,
const SerieBase& serie,
const String& name) const override;
Serie<double> serie(
const Dataframe& dataframe,
DecompDimension targetDim,
const std::string& name) const override;
private:
std::vector<std::string> coordNames_;
};
}Naming Convention
Coordinates uses underscore separation: {name}_{coord}
| Type | Default Names |
|---|---|
Vector2D | {name}_x, {name}_y |
Vector3D | {name}_x, {name}_y, {name}_z |
Vector4D | {name}_x, {name}_y, {name}_z, {name}_w (if 4 names provided) |
Only vector types (Vector2D, Vector3D, Vector4D) are supported.
Matrix and tensor types are ignored by this decomposer.
Examples
Default Coordinate Names
using namespace df;
Serie<Vector3D> positions = {{1,2,3}, {4,5,6}, {7,8,9}};
Dataframe df;
df.add("P", positions);
attributes::Manager mgr(df);
mgr.addDecomposer(attributes::Coordinates());
// Scalar decomposition: "P_x", "P_y", "P_z"
auto names = mgr.getNames(attributes::DecompDimension::Scalar);
auto px = mgr.getSerie<double>("P_x"); // {1, 4, 7}
auto py = mgr.getSerie<double>("P_y"); // {2, 5, 8}
auto pz = mgr.getSerie<double>("P_z"); // {3, 6, 9}
Custom Coordinate Names
using namespace df;
Serie<Vector3D> positions = {{1,2,3}, {4,5,6}};
Dataframe df;
df.add("P", positions);
// Use custom names (e.g., geographic coordinates)
attributes::Manager mgr(df);
mgr.addDecomposer(attributes::Coordinates({"easting", "northing", "elevation"}));
// Scalar decomposition: "P_easting", "P_northing", "P_elevation"
auto names = mgr.getNames(attributes::DecompDimension::Scalar);
auto easting = mgr.getSerie<double>("P_easting"); // {1, 4}
auto northing = mgr.getSerie<double>("P_northing"); // {2, 5}
auto elev = mgr.getSerie<double>("P_elevation"); // {3, 6}
Combining with Components
using namespace df;
Serie<Vector3D> positions = {{1,2,3}, {4,5,6}};
Serie<Stress3D> stresses = {{100,20,10,80,15,60}, {110,25,12,85,18,65}};
Dataframe df;
df.add("P", positions);
df.add("S", stresses);
// Register both decomposers
attributes::Manager mgr(df);
mgr.addDecomposer(attributes::Components());
mgr.addDecomposer(attributes::Coordinates());
auto names = mgr.getNames(attributes::DecompDimension::Scalar);
// Components gives: "Px", "Py", "Pz", "Sxx", "Sxy", ...
// Coordinates gives: "P_x", "P_y", "P_z"
// (Coordinates ignores stress tensors)
// Both decomposers provide position X, with different name patterns
auto px1 = mgr.getSerie<double>("Px"); // via Components
auto px2 = mgr.getSerie<double>("P_x"); // via Coordinates
// Both return the same data: {1, 4}Coordinates vs Components
| Feature | Coordinates | Components |
|---|---|---|
| Supported types | Vectors only | Vectors, matrices, tensors |
| Naming pattern | {name}_{coord} | {name}{suffix} (no separator) |
| Custom names | Yes (constructor parameter) | No (fixed conventions) |
| Use case | Spatial/geographic coordinates | General component extraction |