Difference between revisions of "cpp/numeric/valarray/gslice"
m (Text replace - "{{example cpp" to "{{example") |
m (Text replace - "{{cpp|" to "{{c|") |
||
Line 5: | Line 5: | ||
}} | }} | ||
− | {{tt|std::gslice}} is the selector class that identifies a subset of {{ | + | {{tt|std::gslice}} is the selector class that identifies a subset of {{c|std::valarray}} indices defined by a multi-level set of strides and sizes. Objects of type {{tt|std::gslice}} can be used as indices with valarray's {{tt|operator[]}} to select, for example, columns of a multidimensional array represented as a {{tt|valarray}}. |
Given the starting value {{math|s}}, a list of strides {{math|i{{su|b=j}}}} and a list of sizes {{math|d{{su|b=j}}}}, a {{tt|std::gslice}} constructed from these values selects the set of indices {{math|k{{su|b=j}}{{=}}s+Σ{{su|b=j}}(i{{su|b=j}}d{{su|b=j}})}}. | Given the starting value {{math|s}}, a list of strides {{math|i{{su|b=j}}}} and a list of sizes {{math|d{{su|b=j}}}}, a {{tt|std::gslice}} constructed from these values selects the set of indices {{math|k{{su|b=j}}{{=}}s+Σ{{su|b=j}}(i{{su|b=j}}d{{su|b=j}})}}. | ||
For example, a gslice with starting index {{tt|3}}, strides {{tt|{19,4,1}}} and lengths {{tt|{2,4,3} }} generates the following set of indices: | For example, a gslice with starting index {{tt|3}}, strides {{tt|{19,4,1}}} and lengths {{tt|{2,4,3} }} generates the following set of indices: | ||
− | {{ | + | {{c|1= |
3 + 0*19 + 0*4 + 0*1 = 3, | 3 + 0*19 + 0*4 + 0*1 = 3, | ||
3 + 0*19 + 0*4 + 1*1 = 4, | 3 + 0*19 + 0*4 + 1*1 = 4, |
Revision as of 20:40, 19 April 2012
Template:cpp/numeric/valarray/sidebar
Defined in header <valarray>
|
||
class gslice; |
||
std::gslice
is the selector class that identifies a subset of std::valarray indices defined by a multi-level set of strides and sizes. Objects of type std::gslice
can be used as indices with valarray's operator[]
to select, for example, columns of a multidimensional array represented as a valarray
.
Given the starting value s, a list of strides ij and a list of sizes dj, a std::gslice
constructed from these values selects the set of indices kj=s+Σj(ijdj).
For example, a gslice with starting index 3
, strides {19,4,1
} and lengths {2,4,3}
generates the following set of indices:
3 + 0*19 + 0*4 + 0*1 = 3,
3 + 0*19 + 0*4 + 1*1 = 4,
3 + 0*19 + 0*4 + 2*1 = 5,
3 + 0*19 + 1*4 + 0*1 = 7,
3 + 0*19 + 1*4 + 1*1 = 8,
...
3 + 1*19 + 3*4 + 2*1 = 36
It is possible to construct std::gslice
objects that select some indices more than once: if the above example used the strides {1,1,1}
, the indices would have been {3, 4, 5, 4, 5, 6, ...}
. Such gslices may only be used as arguments to the const version of std::valarray::operator[]
, otherwise the behavior is undefined.
Member functions
constructs a gslice (public member function) | |
accesses the start of the gslice (public member function) | |
accesses the array of strides of the gslice (public member function) | |
accesses the array of sizees of the gslice (public member function) |
Example
demonstrates the use of gslices to address columns of a 3D array
#include <iostream> #include <valarray> void test_print(std::valarray<int>& v, int rows, int cols, int planes) { for(int r=0; r<rows; ++r) { for(int c=0; c<cols; ++c) { for(int z=0; z<planes; ++z) std::cout << v[r*cols*planes + c*planes + z] << ' '; std::cout << '\n'; } std::cout << '\n'; } } int main() { std::valarray<int> v = // 3d array: 2 x 4 x 3 elements { 111,112,113 , 121,122,123 , 131,132,133 , 141,142,143, 211,212,213 , 221,222,223 , 231,232,233 , 241,242,243}; // int ar3d[2][4][3] std::cout << "Initial 2x4x3 array:\n"; test_print(v, 2, 4, 3); // update every value in the first columns of both planes v[std::gslice(0, {2, 4}, {4*3, 3})] = 1; // two level one strides of 12 elements // then four level two strides of 3 elements // subtract the third column from the second column in the 1st plane v[std::gslice(1, {1, 4}, {4*3, 3})] -= v[std::gslice(2, {1, 4}, {4*3, 3})]; std::cout << "After column operations: \n"; test_print(v, 2, 4, 3); }
Output:
Initial 2x4x3 array: 111 112 113 121 122 123 131 132 133 141 142 143 211 212 213 221 222 223 231 232 233 241 242 243 After column operations: 1 -1 113 1 -1 123 1 -1 133 1 -1 143 1 212 213 1 222 223 1 232 233 1 242 243