• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1namespace Eigen {
2
3/** \page TutorialBlockOperations Tutorial page 4 - %Block operations
4    \ingroup Tutorial
5
6\li \b Previous: \ref TutorialArrayClass
7\li \b Next: \ref TutorialAdvancedInitialization
8
9This tutorial page explains the essentials of block operations.
10A block is a rectangular part of a matrix or array. Blocks expressions can be used both
11as rvalues and as lvalues. As usual with Eigen expressions, this abstraction has zero runtime cost
12provided that you let your compiler optimize.
13
14\b Table \b of \b contents
15  - \ref TutorialBlockOperationsUsing
16  - \ref TutorialBlockOperationsSyntaxColumnRows
17  - \ref TutorialBlockOperationsSyntaxCorners
18  - \ref TutorialBlockOperationsSyntaxVectors
19
20
21\section TutorialBlockOperationsUsing Using block operations
22
23The most general block operation in Eigen is called \link DenseBase::block() .block() \endlink.
24There are two versions, whose syntax is as follows:
25
26<table class="manual">
27<tr><th>\b %Block \b operation</td>
28<th>Version constructing a \n dynamic-size block expression</th>
29<th>Version constructing a \n fixed-size block expression</th></tr>
30<tr><td>%Block of size <tt>(p,q)</tt>, starting at <tt>(i,j)</tt></td>
31    <td>\code
32matrix.block(i,j,p,q);\endcode </td>
33    <td>\code
34matrix.block<p,q>(i,j);\endcode </td>
35</tr>
36</table>
37
38As always in Eigen, indices start at 0.
39
40Both versions can be used on fixed-size and dynamic-size matrices and arrays.
41These two expressions are semantically equivalent.
42The only difference is that the fixed-size version will typically give you faster code if the block size is small,
43but requires this size to be known at compile time.
44
45The following program uses the dynamic-size and fixed-size versions to print the values of several blocks inside a
46matrix.
47
48<table class="example">
49<tr><th>Example:</th><th>Output:</th></tr>
50<tr><td>
51\include Tutorial_BlockOperations_print_block.cpp
52</td>
53<td>
54\verbinclude Tutorial_BlockOperations_print_block.out
55</td></tr></table>
56
57In the above example the \link DenseBase::block() .block() \endlink function was employed as a \em rvalue, i.e.
58it was only read from. However, blocks can also be used as \em lvalues, meaning that you can assign to a block.
59
60This is illustrated in the following example. This example also demonstrates blocks in arrays, which works exactly like the above-demonstrated blocks in matrices.
61
62<table class="example">
63<tr><th>Example:</th><th>Output:</th></tr>
64<tr><td>
65\include Tutorial_BlockOperations_block_assignment.cpp
66</td>
67<td>
68\verbinclude Tutorial_BlockOperations_block_assignment.out
69</td></tr></table>
70
71While the \link DenseBase::block() .block() \endlink method can be used for any block operation, there are
72other methods for special cases, providing more specialized API and/or better performance. On the topic of performance, all what
73matters is that you give Eigen as much information as possible at compile time. For example, if your block is a single whole column in a matrix,
74using the specialized \link DenseBase::col() .col() \endlink function described below lets Eigen know that, which can give it optimization opportunities.
75
76The rest of this page describes these specialized methods.
77
78\section TutorialBlockOperationsSyntaxColumnRows Columns and rows
79
80Individual columns and rows are special cases of blocks. Eigen provides methods to easily address them:
81\link DenseBase::col() .col() \endlink and \link DenseBase::row() .row()\endlink.
82
83<table class="manual">
84<tr><th>%Block operation</th>
85<th>Method</th>
86<tr><td>i<sup>th</sup> row
87                    \link DenseBase::row() * \endlink</td>
88    <td>\code
89matrix.row(i);\endcode </td>
90</tr>
91<tr><td>j<sup>th</sup> column
92                    \link DenseBase::col() * \endlink</td>
93    <td>\code
94matrix.col(j);\endcode </td>
95</tr>
96</table>
97
98The argument for \p col() and \p row() is the index of the column or row to be accessed. As always in Eigen, indices start at 0.
99
100<table class="example">
101<tr><th>Example:</th><th>Output:</th></tr>
102<tr><td>
103\include Tutorial_BlockOperations_colrow.cpp
104</td>
105<td>
106\verbinclude Tutorial_BlockOperations_colrow.out
107</td></tr></table>
108
109That example also demonstrates that block expressions (here columns) can be used in arithmetic like any other expression.
110
111
112\section TutorialBlockOperationsSyntaxCorners Corner-related operations
113
114Eigen also provides special methods for blocks that are flushed against one of the corners or sides of a
115matrix or array. For instance, \link DenseBase::topLeftCorner() .topLeftCorner() \endlink can be used to refer
116to a block in the top-left corner of a matrix.
117
118The different possibilities are summarized in the following table:
119
120<table class="manual">
121<tr><th>%Block \b operation</td>
122<th>Version constructing a \n dynamic-size block expression</th>
123<th>Version constructing a \n fixed-size block expression</th></tr>
124<tr><td>Top-left p by q block \link DenseBase::topLeftCorner() * \endlink</td>
125    <td>\code
126matrix.topLeftCorner(p,q);\endcode </td>
127    <td>\code
128matrix.topLeftCorner<p,q>();\endcode </td>
129</tr>
130<tr><td>Bottom-left p by q block
131              \link DenseBase::bottomLeftCorner() * \endlink</td>
132    <td>\code
133matrix.bottomLeftCorner(p,q);\endcode </td>
134    <td>\code
135matrix.bottomLeftCorner<p,q>();\endcode </td>
136</tr>
137<tr><td>Top-right p by q block
138              \link DenseBase::topRightCorner() * \endlink</td>
139    <td>\code
140matrix.topRightCorner(p,q);\endcode </td>
141    <td>\code
142matrix.topRightCorner<p,q>();\endcode </td>
143</tr>
144<tr><td>Bottom-right p by q block
145               \link DenseBase::bottomRightCorner() * \endlink</td>
146    <td>\code
147matrix.bottomRightCorner(p,q);\endcode </td>
148    <td>\code
149matrix.bottomRightCorner<p,q>();\endcode </td>
150</tr>
151<tr><td>%Block containing the first q rows
152                   \link DenseBase::topRows() * \endlink</td>
153    <td>\code
154matrix.topRows(q);\endcode </td>
155    <td>\code
156matrix.topRows<q>();\endcode </td>
157</tr>
158<tr><td>%Block containing the last q rows
159                    \link DenseBase::bottomRows() * \endlink</td>
160    <td>\code
161matrix.bottomRows(q);\endcode </td>
162    <td>\code
163matrix.bottomRows<q>();\endcode </td>
164</tr>
165<tr><td>%Block containing the first p columns
166                    \link DenseBase::leftCols() * \endlink</td>
167    <td>\code
168matrix.leftCols(p);\endcode </td>
169    <td>\code
170matrix.leftCols<p>();\endcode </td>
171</tr>
172<tr><td>%Block containing the last q columns
173                    \link DenseBase::rightCols() * \endlink</td>
174    <td>\code
175matrix.rightCols(q);\endcode </td>
176    <td>\code
177matrix.rightCols<q>();\endcode </td>
178</tr>
179</table>
180
181Here is a simple example illustrating the use of the operations presented above:
182
183<table class="example">
184<tr><th>Example:</th><th>Output:</th></tr>
185<tr><td>
186\include Tutorial_BlockOperations_corner.cpp
187</td>
188<td>
189\verbinclude Tutorial_BlockOperations_corner.out
190</td></tr></table>
191
192
193\section TutorialBlockOperationsSyntaxVectors Block operations for vectors
194
195Eigen also provides a set of block operations designed specifically for the special case of vectors and one-dimensional arrays:
196
197<table class="manual">
198<tr><th> %Block operation</th>
199<th>Version constructing a \n dynamic-size block expression</th>
200<th>Version constructing a \n fixed-size block expression</th></tr>
201<tr><td>%Block containing the first \p n elements
202                    \link DenseBase::head() * \endlink</td>
203    <td>\code
204vector.head(n);\endcode </td>
205    <td>\code
206vector.head<n>();\endcode </td>
207</tr>
208<tr><td>%Block containing the last \p n elements
209                    \link DenseBase::tail() * \endlink</td>
210    <td>\code
211vector.tail(n);\endcode </td>
212    <td>\code
213vector.tail<n>();\endcode </td>
214</tr>
215<tr><td>%Block containing \p n elements, starting at position \p i
216                    \link DenseBase::segment() * \endlink</td>
217    <td>\code
218vector.segment(i,n);\endcode </td>
219    <td>\code
220vector.segment<n>(i);\endcode </td>
221</tr>
222</table>
223
224
225An example is presented below:
226<table class="example">
227<tr><th>Example:</th><th>Output:</th></tr>
228<tr><td>
229\include Tutorial_BlockOperations_vector.cpp
230</td>
231<td>
232\verbinclude Tutorial_BlockOperations_vector.out
233</td></tr></table>
234
235\li \b Next: \ref TutorialAdvancedInitialization
236
237*/
238
239}
240