class TMatrixTSparse<float>: public TMatrixTBase<float>

 Allocate new matrix. Arguments are number of rows, columns, row lowerbound (0 default)
 and column lowerbound (0 default), 0 initialization flag and number of non-zero
 elements (only relevant for sparse format).

 Insert in row rown, n elements of array v at column coln

 Store in array v, n matrix elements of row rown starting at column coln

 General matrix multiplication. Create a matrix C such that C = A * B'.
 Note, matrix C is allocated for constr=1.

 General matrix multiplication. Create a matrix C such that C = A * B'.
 Note, matrix C is allocated for constr=1.

 General matrix multiplication. Create a matrix C such that C = A * B'.
 Note, matrix C is allocated for constr=1.

 General matrix addition. Create a matrix C such that C = A + B.
 Note, matrix C is allocated for constr=1.

 General matrix addition. Create a matrix C such that C = A + B.
 Note, matrix C is allocated for constr=1.

 General matrix subtraction. Create a matrix C such that C = A - B.
 Note, matrix C is allocated for constr=1.

 General matrix subtraction. Create a matrix C such that C = A - B.
 Note, matrix C is allocated for constr=1.

 General matrix subtraction. Create a matrix C such that C = A - B.
 Note, matrix C is allocated for constr=1.

 Copy matrix data to array . It is assumed that array is of size >= fNelems

 Copy nr elements from row/col index and data array to matrix . It is assumed
 that arrays are of size >= nr
 Note that the input arrays are not passed as const since they will be modified !

 Increase/decrease the number of non-zero elements to nelems_new

 Use non-zero data of matrix source to set the sparse structure

 Set the row/column indices to the "sum" of matrices a and b
 It is checked that enough space has been allocated

 Set the row/column indices to the "sum" of matrices a and b
 It is checked that enough space has been allocated

 Set size of the matrix to nrows x ncols with nr_nonzeros non-zero entries
 if nr_nonzeros > 0 .
 New dynamic elements are created, the overlapping part of the old ones are
 copied to the new structures, then the old elements are deleted.

 Set size of the matrix to [row_lwb:row_upb] x [col_lwb:col_upb] with nr_nonzeros
 non-zero entries if nr_nonzeros > 0 .
 New dynamic elemenst are created, the overlapping part of the old ones are
 copied to the new structures, then the old elements are deleted.

 Get submatrix [row_lwb..row_upb][col_lwb..col_upb]; The indexing range of the
 returned matrix depends on the argument option:

 option == "S" : return [0..row_upb-row_lwb+1][0..col_upb-col_lwb+1] (default)
 else          : return [row_lwb..row_upb][col_lwb..col_upb]

 Insert matrix source starting at [row_lwb][col_lwb], thereby overwriting the part
 [row_lwb..row_lwb+nrows_source-1][col_lwb..col_lwb+ncols_source-1];

 Transpose a matrix.

 Make a unit matrix (matrix need not be a square one).

 Row matrix norm, MAX{ SUM{ |M(i,j)|, over j}, over i}.
 The norm is induced by the infinity vector norm.

 Column matrix norm, MAX{ SUM{ |M(i,j)|, over i}, over j}.
 The norm is induced by the 1 vector norm.

 Notice that the sparsity of the matrix is NOT changed : its fRowIndex/fColIndex
 are used !

 Notice that the sparsity of the matrix is NOT changed : its fRowIndex/fColIndex
 are used !

 Assign val to every element of the matrix. Check that the row/col
 indices are set !

 Add val to every element of the matrix.

 Subtract val from every element of the matrix.

 Multiply every element of the matrix with val.

 randomize matrix element values

 randomize matrix element values but keep matrix symmetric positive definite

 Stream an object of class TMatrixTSparse.

 Elementary constructors

 or as a[i][j]