2010-06-26 22:59:18 +08:00
namespace Eigen {
/** \page QuickRefPage Quick reference guide
\b Table \b of \b contents
- \ref QuickRef_Headers
- \ref QuickRef_Types
- \ref QuickRef_Map
- \ref QuickRef_ArithmeticOperators
- \ref QuickRef_Coeffwise
2010-06-27 00:49:50 +08:00
- \ref QuickRef_Reductions
- \ref QuickRef_Blocks
- \ref QuickRef_DiagTriSymm
2010-06-26 22:59:18 +08:00
\n
<hr>
<a href="#" class="top">top</a>
\section QuickRef_Headers Modules and Header files
2010-07-06 20:10:08 +08:00
The Eigen library is divided in a Core module and several additional modules. Each module has a corresponding header file which has to be included in order to use the module. The \c %Dense and \c Eigen header files are provided to conveniently gain access to several modules at once.
2010-06-26 22:59:18 +08:00
<table class="tutorial_code">
<tr><td>Module</td><td>Header file</td><td>Contents</td></tr>
2010-07-06 20:10:08 +08:00
<tr><td>\link Core_Module Core \endlink</td><td>\code#include <Eigen/Core>\endcode</td><td>Matrix and Array classes, basic linear algebra (including triangular and selfadjoint products), array manipulation</td></tr>
<tr><td>\link Geometry_Module Geometry \endlink</td><td>\code#include <Eigen/Geometry>\endcode</td><td>Transform, Translation, Scaling, Rotation2D and 3D rotations (Quaternion, AngleAxis)</td></tr>
<tr><td>\link LU_Module LU \endlink</td><td>\code#include <Eigen/LU>\endcode</td><td>Inverse, determinant, LU decompositions with solver (FullPivLU, PartialPivLU)</td></tr>
<tr><td>\link Cholesky_Module Cholesky \endlink</td><td>\code#include <Eigen/Cholesky>\endcode</td><td>LLT and LDLT Cholesky factorization with solver</td></tr>
<tr><td>\link Householder_Module Householder \endlink</td><td>\code#include <Eigen/Householder>\endcode</td><td>Householder transformations; this module is used by several linear algebra modules</td></tr>
<tr><td>\link SVD_Module SVD \endlink</td><td>\code#include <Eigen/SVD>\endcode</td><td>%SVD decomposition with solver (SVD, JacobiSVD)</td></tr>
<tr><td>\link QR_Module QR \endlink</td><td>\code#include <Eigen/QR>\endcode</td><td>QR decomposition with solver (HouseholderQR, ColPivHouseholderQR, FullPivHouseholderQR)</td></tr>
<tr><td>\link Eigenvalues_Module Eigenvalues \endlink</td><td>\code#include <Eigen/Eigenvalues>\endcode</td><td>Eigenvalue, eigenvector decompositions (EigenSolver, SelfAdjointEigenSolver, ComplexEigenSolver)</td></tr>
<tr><td>\link Sparse_Module Sparse \endlink</td><td>\code#include <Eigen/Sparse>\endcode</td><td>%Sparse matrix storage and related basic linear algebra (SparseMatrix, DynamicSparseMatrix, SparseVector)</td></tr>
<tr><td></td><td>\code#include <Eigen/Dense>\endcode</td><td>Includes Core, Geometry, LU, Cholesky, %SVD, QR, and Eigenvalues header files</td></tr>
<tr><td></td><td>\code#include <Eigen/Eigen>\endcode</td><td>Includes %Dense and %Sparse header files (the whole Eigen library)</td></tr>
2010-06-26 22:59:18 +08:00
</table>
<a href="#" class="top">top</a>
\section QuickRef_Types Array, matrix and vector types
2010-06-28 19:30:10 +08:00
2010-06-26 22:59:18 +08:00
\b Recall: Eigen provides two kinds of dense objects: mathematical matrices and vectors which are both represented by the template class Matrix, and general 1D and 2D arrays represented by the template class Array:
2010-07-01 00:27:27 +08:00
<div class="desired_tutorial_width">
2010-06-26 22:59:18 +08:00
\code
typedef Matrix<Scalar, RowsAtCompileTime, ColsAtCompileTime, Options> MyMatrixType;
typedef Array<Scalar, RowsAtCompileTime, ColsAtCompileTime, Options> MyArrayType;
\endcode
2010-07-01 00:27:27 +08:00
</div>
2010-06-26 22:59:18 +08:00
\li \c Scalar is the scalar type of the coefficients (e.g., \c float, \c double, \c bool, \c int, etc.).
\li \c RowsAtCompileTime and \c ColsAtCompileTime are the number of rows and columns of the matrix as known at compile-time or \c Dynamic.
\li \c Options can be \c ColMajor or \c RowMajor, default is \c ColMajor. (see class Matrix for more options)
All combinations are allowed: you can have a matrix with a fixed number of rows and a dynamic number of columns, etc. The following are all valid:
2010-07-01 00:27:27 +08:00
<div class="desired_tutorial_width">
2010-06-26 22:59:18 +08:00
\code
Matrix<double, 6, Dynamic> // Dynamic number of columns (heap allocation)
Matrix<double, Dynamic, 2> // Dynamic number of rows (heap allocation)
Matrix<double, Dynamic, Dynamic, RowMajor> // Fully dynamic, row major (heap allocation)
Matrix<double, 13, 3> // Fully fixed (static allocation)
\endcode
2010-07-01 00:27:27 +08:00
</div>
2010-06-26 22:59:18 +08:00
In most cases, you can simply use one of the convenience typedefs for \ref matrixtypedefs "matrices" and \ref arraytypedefs "arrays". Some examples:
<table class="tutorial_code">
<tr><td>\code
Matrix<float,Dynamic,Dynamic> <=> MatrixXf
Matrix<double,Dynamic,1> <=> VectorXd
Matrix<int,1,Dynamic> <=> RowVectorXi
Matrix<float,3,3> <=> Matrix3f
Matrix<float,4,1> <=> Vector4f
\endcode</td><td>\code
Array<float,Dynamic,Dynamic> <=> ArrayXXf
Array<double,Dynamic,1> <=> ArrayXd
Array<int,1,Dynamic> <=> RowArrayXi
Array<float,3,3> <=> Array33f
Array<float,4,1> <=> Array4f
\endcode</td></tr>
</table>
Conversion between the matrix and array worlds:
2010-07-01 00:27:27 +08:00
<div class="desired_tutorial_width">
2010-06-26 22:59:18 +08:00
\code
Array44f a1, a1;
Matrix4f m1, m2;
2010-06-27 04:19:03 +08:00
m1 = a1 * a2; // coeffwise product, implicit conversion from array to matrix.
a1 = m1 * m2; // matrix product, implicit conversion from matrix to array.
a2 = a1 + m1.array(); // mixing array and matrix is forbidden
m2 = a1.matrix() + m1; // and explicit conversion is required.
ArrayWrapper<Matrix4f> m1a(m1); // m1a is an alias for m1.array(), they share the same coefficients
2010-06-26 22:59:18 +08:00
MatrixWrapper<Array44f> a1m(a1);
\endcode
2010-07-01 00:27:27 +08:00
</div>
2010-06-26 22:59:18 +08:00
In the rest of this document we will use the following symbols to emphasize the features which are specifics to a given kind of object:
\li <a name="matrixonly"><a/>\matrixworld linear algebra matrix and vector only
\li <a name="arrayonly"><a/>\arrayworld array objects only
\subsection QuickRef_Basics Basic matrix manipulation
<table class="tutorial_code">
<tr><td></td><td>1D objects</td><td>2D objects</td><td>Notes</td></tr>
<tr><td>Constructors</td>
<td>\code
Vector4d v4;
Vector2f v1(x, y);
Array3i v2(x, y, z);
Vector4d v3(x, y, z, w);
2010-06-27 04:19:03 +08:00
VectorXf v5; // empty object
2010-06-26 22:59:18 +08:00
ArrayXf v6(size);
\endcode</td><td>\code
Matrix4f m1;
2010-06-27 04:19:03 +08:00
MatrixXf m5; // empty object
2010-06-26 22:59:18 +08:00
MatrixXf m6(nb_rows, nb_columns);
2010-06-27 04:19:03 +08:00
\endcode</td><td class="note">
By default, the coefficients \n are left uninitialized</td></tr>
2010-06-26 22:59:18 +08:00
<tr><td>Comma initializer</td>
<td>\code
Vector3f v1; v1 << x, y, z;
ArrayXf v2(4); v2 << 1, 2, 3, 4;
\endcode</td><td>\code
Matrix3f m1; m1 << 1, 2, 3,
4, 5, 6,
7, 8, 9;
\endcode</td><td></td></tr>
<tr><td>Comma initializer (bis)</td>
<td colspan="2">
\include Tutorial_commainit_02.cpp
</td>
<td>
output:
\verbinclude Tutorial_commainit_02.out
2010-07-01 00:04:36 +08:00
</td>
</tr>
2010-06-26 22:59:18 +08:00
<tr><td>Runtime info</td>
<td>\code
vector.size();
vector.innerStride();
vector.data();
\endcode</td><td>\code
matrix.rows(); matrix.cols();
matrix.innerSize(); matrix.outerSize();
matrix.innerStride(); matrix.outerStride();
matrix.data();
2010-06-30 21:52:00 +08:00
\endcode</td><td class="note">Inner/Outer* are storage order dependent</td></tr>
2010-06-26 22:59:18 +08:00
<tr><td>Compile-time info</td>
<td colspan="2">\code
ObjectType::Scalar ObjectType::RowsAtCompileTime
ObjectType::RealScalar ObjectType::ColsAtCompileTime
ObjectType::Index ObjectType::SizeAtCompileTime
\endcode</td><td></td></tr>
<tr><td>Resizing</td>
<td>\code
vector.resize(size);
vector.resizeLike(other_vector);
vector.conservativeResize(size);
\endcode</td><td>\code
matrix.resize(nb_rows, nb_cols);
matrix.resize(Eigen::NoChange, nb_cols);
matrix.resize(nb_rows, Eigen::NoChange);
matrix.resizeLike(other_matrix);
matrix.conservativeResize(nb_rows, nb_cols);
2010-06-30 21:52:00 +08:00
\endcode</td><td class="note">no-op if the new sizes match,<br/>otherwise data are lost<br/><br/>resizing with data preservation</td></tr>
2010-06-26 22:59:18 +08:00
<tr><td>Coeff access with \n range checking</td>
<td>\code
vector(i) vector.x()
vector[i] vector.y()
vector.z()
vector.w()
\endcode</td><td>\code
matrix(i,j)
2010-08-10 18:37:23 +08:00
\endcode</td><td class="note">Range checking is disabled if \n NDEBUG or #EIGEN_NO_DEBUG is defined</td></tr>
2010-06-26 22:59:18 +08:00
<tr><td>Coeff access without \n range checking</td>
<td>\code
vector.coeff(i)
vector.coeffRef(i)
\endcode</td><td>\code
matrix.coeff(i,j)
matrix.coeffRef(i,j)
\endcode</td><td></td></tr>
<tr><td>Assignment/copy</td>
<td colspan="2">\code
object = expression;
object_of_float = expression_of_double.cast<float>();
2010-06-27 04:19:03 +08:00
\endcode</td><td class="note">the destination is automatically resized (if possible)</td></tr>
2010-06-26 22:59:18 +08:00
</table>
\subsection QuickRef_PredefMat Predefined Matrices
<table class="tutorial_code">
<tr>
<td>Fixed-size matrix or vector</td>
<td>Dynamic-size matrix</td>
<td>Dynamic-size vector</td>
</tr>
<tr style="border-bottom-style: none;">
<td>
\code
typedef {Matrix3f|Array33f} FixedXD;
FixedXD x;
x = FixedXD::Zero();
x = FixedXD::Ones();
x = FixedXD::Constant(value);
x = FixedXD::Random();
2010-07-21 04:55:22 +08:00
x = FixedXD::LinSpaced(low, high, size);
2010-06-26 22:59:18 +08:00
x.setZero();
x.setOnes();
x.setConstant(value);
x.setRandom();
2010-07-21 04:55:22 +08:00
x.setLinSpaced(low, high, size);
2010-06-26 22:59:18 +08:00
\endcode
</td>
<td>
\code
typedef {MatrixXf|ArrayXXf} Dynamic2D;
Dynamic2D x;
x = Dynamic2D::Zero(rows, cols);
x = Dynamic2D::Ones(rows, cols);
x = Dynamic2D::Constant(rows, cols, value);
x = Dynamic2D::Random(rows, cols);
2010-07-21 04:55:22 +08:00
N/A
2010-06-26 22:59:18 +08:00
x.setZero(rows, cols);
x.setOnes(rows, cols);
x.setConstant(rows, cols, value);
x.setRandom(rows, cols);
2010-07-21 04:55:22 +08:00
N/A
2010-06-26 22:59:18 +08:00
\endcode
</td>
<td>
\code
typedef {VectorXf|ArrayXf} Dynamic1D;
Dynamic1D x;
x = Dynamic1D::Zero(size);
x = Dynamic1D::Ones(size);
x = Dynamic1D::Constant(size, value);
x = Dynamic1D::Random(size);
2010-07-21 04:55:22 +08:00
x = Dynamic1D::LinSpaced(low, high, size);
2010-06-26 22:59:18 +08:00
x.setZero(size);
x.setOnes(size);
x.setConstant(size, value);
x.setRandom(size);
2010-07-21 04:55:22 +08:00
x.setLinSpaced(low, high, size);
2010-06-26 22:59:18 +08:00
\endcode
</td>
</tr>
<tr><td colspan="3">Identity and \link MatrixBase::Unit basis vectors \endlink \matrixworld</td></tr>
<tr style="border-bottom-style: none;">
<td>
\code
x = FixedXD::Identity();
x.setIdentity();
Vector3f::UnitX() // 1 0 0
Vector3f::UnitY() // 0 1 0
Vector3f::UnitZ() // 0 0 1
\endcode
</td>
<td>
\code
x = Dynamic2D::Identity(rows, cols);
x.setIdentity(rows, cols);
N/A
\endcode
</td>
<td>\code
N/A
VectorXf::Unit(size,i)
VectorXf::Unit(4,1) == Vector4f(0,1,0,0)
== Vector4f::UnitY()
\endcode
</td>
</tr>
</table>
2010-06-27 04:19:03 +08:00
\subsection QuickRef_Map Mapping external arrays
2010-06-26 22:59:18 +08:00
<table class="tutorial_code">
<tr>
2010-06-27 04:19:03 +08:00
<td>Contiguous \n memory</td>
2010-06-26 22:59:18 +08:00
<td>\code
float data[] = {1,2,3,4};
2010-06-27 04:19:03 +08:00
Map<Vector3f> v1(data); // uses v1 as a Vector3f object
Map<ArrayXf> v2(data,3); // uses v2 as a ArrayXf object
Map<Array22f> m1(data); // uses m1 as a Array22f object
Map<MatrixXf> m2(data,2,2); // uses m2 as a MatrixXf object
2010-06-26 22:59:18 +08:00
\endcode</td>
</tr>
<tr>
2010-06-27 04:19:03 +08:00
<td>Typical usage \n of strides</td>
2010-06-26 22:59:18 +08:00
<td>\code
float data[] = {1,2,3,4,5,6,7,8,9};
2010-06-27 04:19:03 +08:00
Map<VectorXf,0,InnerStride<2> > v1(data,3); // = [1,3,5]
Map<VectorXf,0,InnerStride<> > v2(data,3,InnerStride<>(3)); // = [1,4,7]
Map<MatrixXf,0,OuterStride<3> > m2(data,2,3); // both lines |1,4,7|
Map<MatrixXf,0,OuterStride<> > m1(data,2,3,OuterStride<>(3)); // are equal to: |2,5,8|
2010-06-26 22:59:18 +08:00
\endcode</td>
</tr>
</table>
<a href="#" class="top">top</a>
\section QuickRef_ArithmeticOperators Arithmetic Operators
<table class="tutorial_code">
<tr><td>
2010-06-27 04:19:03 +08:00
add \n subtract</td><td>\code
mat3 = mat1 + mat2; mat3 += mat1;
mat3 = mat1 - mat2; mat3 -= mat1;\endcode
2010-06-26 22:59:18 +08:00
</td></tr>
<tr><td>
scalar product</td><td>\code
2010-06-27 04:19:03 +08:00
mat3 = mat1 * s1; mat3 *= s1; mat3 = s1 * mat1;
mat3 = mat1 / s1; mat3 /= s1;\endcode
2010-06-26 22:59:18 +08:00
</td></tr>
<tr><td>
2010-06-27 04:19:03 +08:00
matrix/vector \n products \matrixworld</td><td>\code
2010-06-26 22:59:18 +08:00
col2 = mat1 * col1;
2010-06-27 04:19:03 +08:00
row2 = row1 * mat1; row1 *= mat1;
mat3 = mat1 * mat2; mat3 *= mat1; \endcode
2010-06-26 22:59:18 +08:00
</td></tr>
<tr><td>
2010-06-27 04:19:03 +08:00
transposition \n adjoint \matrixworld</td><td>\code
2010-06-27 00:49:50 +08:00
mat1 = mat2.transpose(); mat1.transposeInPlace();
mat1 = mat2.adjoint(); mat1.adjointInPlace();
\endcode
</td></tr>
<tr><td>
2010-06-27 04:19:03 +08:00
\link MatrixBase::dot() dot \endlink product \n inner product \matrixworld</td><td>\code
scalar = vec1.dot(vec2);
2010-06-26 22:59:18 +08:00
scalar = col1.adjoint() * col2;
2010-06-27 04:19:03 +08:00
scalar = (col1.adjoint() * col2).value();\endcode
2010-06-26 22:59:18 +08:00
</td></tr>
<tr><td>
outer product \matrixworld</td><td>\code
mat = col1 * col2.transpose();\endcode
</td></tr>
2010-06-27 00:49:50 +08:00
<tr><td>
2010-06-27 04:19:03 +08:00
\link MatrixBase::norm() norm \endlink \n \link MatrixBase::normalized() normalization \endlink \matrixworld</td><td>\code
2010-06-27 00:49:50 +08:00
scalar = vec1.norm(); scalar = vec1.squaredNorm()
vec2 = vec1.normalized(); vec1.normalize(); // inplace \endcode
</td></tr>
2010-06-26 22:59:18 +08:00
<tr><td>
\link MatrixBase::cross() cross product \endlink \matrixworld</td><td>\code
#include <Eigen/Geometry>
vec3 = vec1.cross(vec2);\endcode</td></tr>
</table>
<a href="#" class="top">top</a>
\section QuickRef_Coeffwise Coefficient-wise \& Array operators
Coefficient-wise operators for matrices and vectors:
<table class="tutorial_code">
<tr><td>Matrix API \matrixworld</td><td>Via Array conversions</td></tr>
<tr><td>\code
mat1.cwiseMin(mat2)
mat1.cwiseMax(mat2)
2010-06-28 19:30:10 +08:00
mat1.cwiseAbs2()
2010-06-26 22:59:18 +08:00
mat1.cwiseAbs()
mat1.cwiseSqrt()
mat1.cwiseProduct(mat2)
mat1.cwiseQuotient(mat2)\endcode
</td><td>\code
mat1.array().min(mat2.array())
mat1.array().max(mat2.array())
mat1.array().abs2()
mat1.array().abs()
mat1.array().sqrt()
mat1.array() * mat2.array()
mat1.array() / mat2.array()
\endcode</td></tr>
</table>
Array operators:\arrayworld
<table class="tutorial_code">
<tr><td>Arithmetic operators</td><td>\code
array1 * array2 array1 / array2 array1 *= array2 array1 /= array2
array1 + scalar array1 - scalar array1 += scalar array1 -= scalar
\endcode</td></tr>
<tr><td>Comparisons</td><td>\code
array1 < array2 array1 > array2 array1 < scalar array1 > scalar
array1 <= array2 array1 >= array2 array1 <= scalar array1 >= scalar
array1 == array2 array1 != array2 array1 == scalar array1 != scalar
\endcode</td></tr>
2010-06-27 04:19:03 +08:00
<tr><td>Trigo, power, and \n misc functions \n and the STL variants</td><td>\code
2010-06-26 22:59:18 +08:00
array1.min(array2) std::min(array1,array2)
array1.max(array2) std::max(array1,array2)
array1.abs2()
array1.abs() std::abs(array1)
array1.sqrt() std::sqrt(array1)
array1.log() std::log(array1)
array1.exp() std::exp(array1)
array1.pow(exponent) std::pow(array1,exponent)
array1.square()
array1.cube()
array1.inverse()
array1.sin() std::sin(array1)
array1.cos() std::cos(array1)
array1.tan() std::tan(array1)
\endcode
</td></tr>
</table>
<a href="#" class="top">top</a>
2010-06-27 00:49:50 +08:00
\section QuickRef_Reductions Reductions
2010-06-26 22:59:18 +08:00
Eigen provides several reduction methods such as:
\link DenseBase::minCoeff() minCoeff() \endlink, \link DenseBase::maxCoeff() maxCoeff() \endlink,
2010-06-27 04:19:03 +08:00
\link DenseBase::sum() sum() \endlink, \link DenseBase::prod() prod() \endlink,
\link MatrixBase::trace() trace() \endlink \matrixworld,
2010-06-26 22:59:18 +08:00
\link MatrixBase::norm() norm() \endlink \matrixworld, \link MatrixBase::squaredNorm() squaredNorm() \endlink \matrixworld,
\link DenseBase::all() all() \endlink \redstar,and \link DenseBase::any() any() \endlink \redstar.
All reduction operations can be done matrix-wise,
\link DenseBase::colwise() column-wise \endlink \redstar or
\link DenseBase::rowwise() row-wise \endlink \redstar. Usage example:
<table class="tutorial_code">
<tr><td rowspan="3" style="border-right-style:dashed">\code
5 3 1
mat = 2 7 8
9 4 6 \endcode
</td> <td>\code mat.minCoeff(); \endcode</td><td>\code 1 \endcode</td></tr>
<tr><td>\code mat.colwise().minCoeff(); \endcode</td><td>\code 2 3 1 \endcode</td></tr>
<tr><td>\code mat.rowwise().minCoeff(); \endcode</td><td>\code
1
2
4
\endcode</td></tr>
</table>
2010-07-26 03:29:07 +08:00
Special versions of \link DenseBase::minCoeff(Index*,Index*) minCoeff \endlink and \link DenseBase::maxCoeff(Index*,Index*) maxCoeff \endlink:
2010-07-01 00:27:27 +08:00
<div class="desired_tutorial_width">
2010-06-27 04:19:03 +08:00
\code
int i, j;
s = vector.minCoeff(&i); // s == vector[i]
s = matrix.maxCoeff(&i, &j); // s == matrix(i,j)
\endcode
2010-07-01 00:27:27 +08:00
</div>
2010-06-27 04:19:03 +08:00
Typical use cases of all() and any():
2010-07-01 00:27:27 +08:00
<div class="desired_tutorial_width">
2010-06-27 04:19:03 +08:00
\code
if((array1 > 0).all()) ... // if all coefficients of array1 are greater than 0 ...
if((array1 < array2).any()) ... // if there exist a pair i,j such that array1(i,j) < array2(i,j) ...
\endcode
2010-07-01 00:27:27 +08:00
</div>
2010-06-26 22:59:18 +08:00
2010-06-27 04:19:03 +08:00
<a href="#" class="top">top</a>\section QuickRef_Blocks Sub-matrices
2010-06-26 22:59:18 +08:00
2010-07-26 03:29:07 +08:00
Read-write access to a \link DenseBase::col(Index) column \endlink
or a \link DenseBase::row(Index) row \endlink of a matrix (or array):
2010-07-01 00:27:27 +08:00
<div class="desired_tutorial_width">
2010-06-26 22:59:18 +08:00
\code
mat1.row(i) = mat2.col(j);
mat1.col(j1).swap(mat1.col(j2));
\endcode
2010-07-01 00:27:27 +08:00
</div>
2010-06-26 22:59:18 +08:00
Read-write access to sub-vectors:
<table class="tutorial_code">
<tr>
<td>Default versions</td>
<td>Optimized versions when the size \n is known at compile time</td></tr>
<td></td>
<tr><td>\code vec1.head(n)\endcode</td><td>\code vec1.head<n>()\endcode</td><td>the first \c n coeffs </td></tr>
<tr><td>\code vec1.tail(n)\endcode</td><td>\code vec1.tail<n>()\endcode</td><td>the last \c n coeffs </td></tr>
<tr><td>\code vec1.segment(pos,n)\endcode</td><td>\code vec1.segment<n>(pos)\endcode</td>
2010-07-06 20:10:08 +08:00
<td>the \c n coeffs in \n the range [\c pos : \c pos + \c n [</td></tr>
2010-06-26 22:59:18 +08:00
<tr style="border-style: dashed none dashed none;"><td>
Read-write access to sub-matrices:</td><td></td><td></td></tr>
<tr>
<td>\code mat1.block(i,j,rows,cols)\endcode
2010-07-26 03:29:07 +08:00
\link DenseBase::block(Index,Index,Index,Index) (more) \endlink</td>
2010-06-26 22:59:18 +08:00
<td>\code mat1.block<rows,cols>(i,j)\endcode
2010-07-26 03:29:07 +08:00
\link DenseBase::block(Index,Index) (more) \endlink</td>
2010-06-27 00:49:50 +08:00
<td>the \c rows x \c cols sub-matrix \n starting from position (\c i,\c j)</td></tr>
<tr><td>\code
2010-06-26 22:59:18 +08:00
mat1.topLeftCorner(rows,cols)
mat1.topRightCorner(rows,cols)
mat1.bottomLeftCorner(rows,cols)
mat1.bottomRightCorner(rows,cols)\endcode
<td>\code
mat1.topLeftCorner<rows,cols>()
mat1.topRightCorner<rows,cols>()
mat1.bottomLeftCorner<rows,cols>()
mat1.bottomRightCorner<rows,cols>()\endcode
<td>the \c rows x \c cols sub-matrix \n taken in one of the four corners</td></tr>
2010-06-27 00:49:50 +08:00
<tr><td>\code
mat1.topRows(rows)
mat1.bottomRows(rows)
mat1.leftCols(cols)
mat1.rightCols(cols)\endcode
2010-06-26 22:59:18 +08:00
<td>\code
2010-06-27 00:49:50 +08:00
mat1.topRows<rows>()
mat1.bottomRows<rows>()
mat1.leftCols<cols>()
mat1.rightCols<cols>()\endcode
2010-06-27 04:19:03 +08:00
<td>specialized versions of block() \n when the block fit two corners</td></tr>
2010-06-26 22:59:18 +08:00
</table>
2010-06-27 00:49:50 +08:00
<a href="#" class="top">top</a>\section QuickRef_DiagTriSymm Diagonal, Triangular, and Self-adjoint matrices
(matrix world \matrixworld)
2010-06-26 22:59:18 +08:00
2010-06-27 00:49:50 +08:00
\subsection QuickRef_Diagonal Diagonal matrices
2010-06-26 22:59:18 +08:00
<table class="tutorial_code">
<tr><td>
2010-06-27 04:19:03 +08:00
view a vector \link MatrixBase::asDiagonal() as a diagonal matrix \endlink \n </td><td>\code
2010-06-27 00:49:50 +08:00
mat1 = vec1.asDiagonal();\endcode
2010-06-26 22:59:18 +08:00
</td></tr>
<tr><td>
2010-06-27 00:49:50 +08:00
Declare a diagonal matrix</td><td>\code
DiagonalMatrix<Scalar,SizeAtCompileTime> diag1(size);
diag1.diagonal() = vector;\endcode
2010-06-26 22:59:18 +08:00
</td></tr>
2010-07-26 03:29:07 +08:00
<tr><td>Access the \link MatrixBase::diagonal() diagonal \endlink and \link MatrixBase::diagonal(Index) super/sub diagonals \endlink of a matrix as a vector (read/write)</td>
2010-06-27 00:49:50 +08:00
<td>\code
2010-06-27 04:19:03 +08:00
vec1 = mat1.diagonal(); mat1.diagonal() = vec1; // main diagonal
vec1 = mat1.diagonal(+n); mat1.diagonal(+n) = vec1; // n-th super diagonal
vec1 = mat1.diagonal(-n); mat1.diagonal(-n) = vec1; // n-th sub diagonal
vec1 = mat1.diagonal<1>(); mat1.diagonal<1>() = vec1; // first super diagonal
vec1 = mat1.diagonal<-2>(); mat1.diagonal<-2>() = vec1; // second sub diagonal
2010-06-27 00:49:50 +08:00
\endcode</td>
</tr>
2010-06-26 22:59:18 +08:00
2010-06-27 00:49:50 +08:00
<tr><td>Optimized products and inverse</td>
<td>\code
mat3 = scalar * diag1 * mat1;
mat3 += scalar * mat1 * vec1.asDiagonal();
mat3 = vec1.asDiagonal().inverse() * mat1
mat3 = mat1 * diag1.inverse()
\endcode</td>
</tr>
2010-06-26 22:59:18 +08:00
2010-06-27 00:49:50 +08:00
</table>
2010-06-26 22:59:18 +08:00
2010-06-27 00:49:50 +08:00
\subsection QuickRef_TriangularView Triangular views
2010-06-26 22:59:18 +08:00
2010-06-27 04:19:03 +08:00
TriangularView gives a view on a triangular part of a dense matrix and allows to perform optimized operations on it. The opposite triangular part is never referenced and can be used to store other information.
2010-06-26 22:59:18 +08:00
<table class="tutorial_code">
<tr><td>
2010-06-27 04:19:03 +08:00
Reference to a triangular with optional \n
unit or null diagonal (read/write):
2010-06-26 22:59:18 +08:00
</td><td>\code
2010-06-27 00:49:50 +08:00
m.triangularView<Xxx>()
\endcode \n
\c Xxx = Upper, Lower, StrictlyUpper, StrictlyLower, UnitUpper, UnitLower
2010-06-26 22:59:18 +08:00
</td></tr>
<tr><td>
Writing to a specific triangular part:\n (only the referenced triangular part is evaluated)
</td><td>\code
2010-06-27 00:49:50 +08:00
m1.triangularView<Eigen::Lower>() = m2 + m3 \endcode
2010-06-26 22:59:18 +08:00
</td></tr>
<tr><td>
Conversion to a dense matrix setting the opposite triangular part to zero:
</td><td>\code
2010-06-27 00:49:50 +08:00
m2 = m1.triangularView<Eigen::UnitUpper>()\endcode
2010-06-26 22:59:18 +08:00
</td></tr>
<tr><td>
Products:
</td><td>\code
2010-06-27 00:49:50 +08:00
m3 += s1 * m1.adjoint().triangularView<Eigen::UnitUpper>() * m2
m3 -= s1 * m2.conjugate() * m1.adjoint().triangularView<Eigen::Lower>() \endcode
2010-06-26 22:59:18 +08:00
</td></tr>
<tr><td>
2010-06-27 04:19:03 +08:00
Solving linear equations:\n
\f$ M_2 := L_1^{-1} M_2 \f$ \n
\f$ M_3 := {L_1^*}^{-1} M_3 \f$ \n
2010-07-22 19:15:15 +08:00
\f$ M_4 := M_4 U_1^{-1} \f$
2010-06-27 04:19:03 +08:00
</td><td>\n \code
L1.triangularView<Eigen::UnitLower>().solveInPlace(M2)
L1.triangularView<Eigen::Lower>().adjoint().solveInPlace(M3)
U1.triangularView<Eigen::Upper>().solveInPlace<OnTheRight>(M4)\endcode
2010-06-26 22:59:18 +08:00
</td></tr>
</table>
2010-06-27 00:49:50 +08:00
\subsection QuickRef_SelfadjointMatrix Symmetric/selfadjoint views
2010-06-26 22:59:18 +08:00
2010-06-27 04:19:03 +08:00
Just as for triangular matrix, you can reference any triangular part of a square matrix to see it as a selfadjoint
matrix and perform special and optimized operations. Again the opposite triangular part is never referenced and can be
2010-06-26 22:59:18 +08:00
used to store other information.
<table class="tutorial_code">
<tr><td>
Conversion to a dense matrix:
</td><td>\code
2010-06-27 00:49:50 +08:00
m2 = m.selfadjointView<Eigen::Lower>();\endcode
2010-06-26 22:59:18 +08:00
</td></tr>
<tr><td>
Product with another general matrix or vector:
</td><td>\code
2010-06-27 00:49:50 +08:00
m3 = s1 * m1.conjugate().selfadjointView<Eigen::Upper>() * m3;
m3 -= s1 * m3.adjoint() * m1.selfadjointView<Eigen::Lower>();\endcode
2010-06-26 22:59:18 +08:00
</td></tr>
<tr><td>
2010-06-27 04:19:03 +08:00
Rank 1 and rank K update: \n
\f$ upper(M_1) += s1 M_2^* M_2 \f$ \n
\f$ lower(M_1) -= M_2 M_2^* \f$
</td><td>\n \code
M1.selfadjointView<Eigen::Upper>().rankUpdate(M2,s1);
2010-06-27 00:49:50 +08:00
m1.selfadjointView<Eigen::Lower>().rankUpdate(m2.adjoint(),-1); \endcode
2010-06-26 22:59:18 +08:00
</td></tr>
<tr><td>
2010-06-27 04:19:03 +08:00
Rank 2 update: (\f$ M += s u v^* + s v u^* \f$)
2010-06-26 22:59:18 +08:00
</td><td>\code
2010-06-27 04:19:03 +08:00
M.selfadjointView<Eigen::Upper>().rankUpdate(u,v,s);
2010-06-26 22:59:18 +08:00
\endcode
</td></tr>
<tr><td>
2010-06-27 04:19:03 +08:00
Solving linear equations:\n(\f$ M_2 := M_1^{-1} M_2 \f$)
2010-06-26 22:59:18 +08:00
</td><td>\code
// via a standard Cholesky factorization
2010-06-27 04:19:03 +08:00
m2 = m1.selfadjointView<Eigen::Upper>().llt().solve(m2);
2010-06-26 22:59:18 +08:00
// via a Cholesky factorization with pivoting
2010-06-27 04:19:03 +08:00
m2 = m1.selfadjointView<Eigen::Lower>().ldlt().solve(m2);
2010-06-26 22:59:18 +08:00
\endcode
</td></tr>
</table>
2010-06-27 00:49:50 +08:00
*/
2010-06-26 22:59:18 +08:00
2010-06-27 00:49:50 +08:00
/*
<table class="tutorial_code">
<tr><td>
\link MatrixBase::asDiagonal() make a diagonal matrix \endlink \n from a vector </td><td>\code
mat1 = vec1.asDiagonal();\endcode
</td></tr>
<tr><td>
Declare a diagonal matrix</td><td>\code
DiagonalMatrix<Scalar,SizeAtCompileTime> diag1(size);
diag1.diagonal() = vector;\endcode
</td></tr>
<tr><td>Access \link MatrixBase::diagonal() the diagonal and super/sub diagonals of a matrix \endlink as a vector (read/write)</td>
<td>\code
vec1 = mat1.diagonal(); mat1.diagonal() = vec1; // main diagonal
vec1 = mat1.diagonal(+n); mat1.diagonal(+n) = vec1; // n-th super diagonal
vec1 = mat1.diagonal(-n); mat1.diagonal(-n) = vec1; // n-th sub diagonal
vec1 = mat1.diagonal<1>(); mat1.diagonal<1>() = vec1; // first super diagonal
vec1 = mat1.diagonal<-2>(); mat1.diagonal<-2>() = vec1; // second sub diagonal
\endcode</td>
</tr>
2010-06-26 22:59:18 +08:00
2010-06-27 00:49:50 +08:00
<tr><td>View on a triangular part of a matrix (read/write)</td>
<td>\code
mat2 = mat1.triangularView<Xxx>();
// Xxx = Upper, Lower, StrictlyUpper, StrictlyLower, UnitUpper, UnitLower
mat1.triangularView<Upper>() = mat2 + mat3; // only the upper part is evaluated and referenced
\endcode</td></tr>
2010-06-26 22:59:18 +08:00
2010-06-27 00:49:50 +08:00
<tr><td>View a triangular part as a symmetric/self-adjoint matrix (read/write)</td>
<td>\code
mat2 = mat1.selfadjointView<Xxx>(); // Xxx = Upper or Lower
mat1.selfadjointView<Upper>() = mat2 + mat2.adjoint(); // evaluated and write to the upper triangular part only
\endcode</td></tr>
2010-06-26 22:59:18 +08:00
2010-06-27 00:49:50 +08:00
</table>
Optimized products:
\code
mat3 += scalar * vec1.asDiagonal() * mat1
mat3 += scalar * mat1 * vec1.asDiagonal()
mat3.noalias() += scalar * mat1.triangularView<Xxx>() * mat2
mat3.noalias() += scalar * mat2 * mat1.triangularView<Xxx>()
mat3.noalias() += scalar * mat1.selfadjointView<Upper or Lower>() * mat2
mat3.noalias() += scalar * mat2 * mat1.selfadjointView<Upper or Lower>()
mat1.selfadjointView<Upper or Lower>().rankUpdate(mat2);
mat1.selfadjointView<Upper or Lower>().rankUpdate(mat2.adjoint(), scalar);
\endcode
Inverse products: (all are optimized)
\code
mat3 = vec1.asDiagonal().inverse() * mat1
mat3 = mat1 * diag1.inverse()
mat1.triangularView<Xxx>().solveInPlace(mat2)
mat1.triangularView<Xxx>().solveInPlace<OnTheRight>(mat2)
mat2 = mat1.selfadjointView<Upper or Lower>().llt().solve(mat2)
\endcode
*/
2010-06-26 22:59:18 +08:00
}