如何在Rcpp中按行写入列表中矩阵与稀疏矩阵的交错结果至文件？

Here's a polished, functional version of your code that handles both dense matrix and sparse dgCMatrix objects, writes rows to a file in the interleaved order you specified, and includes error checking for robustness:

#include <RcppArmadillo.h>
#include <fstream>
#include <vector>

// [[Rcpp::export]]
bool write_rows(Rcpp::List data, Rcpp::CharacterVector clss, int n, std::string filename) {
    // Open output file
    std::ofstream outfile(filename);
    if (!outfile.is_open()) {
        Rcpp::Rcerr << "Error: Could not open file '" << filename << "' for writing." << std::endl;
        return false;
    }

    int num_objects = data.length();
    
    // Validate input lengths match
    if (clss.length() != num_objects) {
        Rcpp::Rcerr << "Error: Length of 'clss' must match length of 'data'." << std::endl;
        outfile.close();
        return false;
    }

    for (int row_idx = 0; row_idx < n; ++row_idx) {
        for (int obj_idx = 0; obj_idx < num_objects; ++obj_idx) {
            std::string current_class = Rcpp::as<std::string>(clss[obj_idx]);

            if (current_class == "matrix") {
                Rcpp::NumericMatrix mat = Rcpp::as<Rcpp::NumericMatrix>(data[obj_idx]);
                
                // Check if row index is within bounds
                if (row_idx >= mat.nrow()) {
                    Rcpp::Rcerr << "Error: Row index " << row_idx << " exceeds rows in matrix at position " << obj_idx << "." << std::endl;
                    outfile.close();
                    return false;
                }

                // Write dense row elements (comma-separated)
                Rcpp::NumericMatrix::Row row = mat.row(row_idx);
                for (int col_idx = 0; col_idx < row.size(); ++col_idx) {
                    if (col_idx > 0) outfile << ",";
                    outfile << row[col_idx];
                }
                outfile << "\n";

            } else if (current_class == "dgCMatrix") {
                arma::sp_mat sp_mat = Rcpp::as<arma::sp_mat>(data[obj_idx]);
                
                // Check if row index is within bounds
                if (row_idx >= sp_mat.n_rows) {
                    Rcpp::Rcerr << "Error: Row index " << row_idx << " exceeds rows in dgCMatrix at position " << obj_idx << "." << std::endl;
                    outfile.close();
                    return false;
                }

                // Convert sparse row to dense vector (preserves zeros)
                std::vector<double> row_data(sp_mat.n_cols, 0.0);
                arma::sp_mat::row_iterator it = sp_mat.begin_row(row_idx);
                arma::sp_mat::row_iterator it_end = sp_mat.end_row(row_idx);

                for (; it != it_end; ++it) {
                    int col_pos = it.col();
                    row_data[col_pos] = *it;
                }

                // Write dense row elements (comma-separated)
                for (int col_idx = 0; col_idx < row_data.size(); ++col_idx) {
                    if (col_idx > 0) outfile << ",";
                    outfile << row_data[col_idx];
                }
                outfile << "\n";

            } else {
                Rcpp::Rcerr << "Error: Unsupported class '" << current_class << "' at position " << obj_idx << "." << std::endl;
                outfile.close();
                return false;
            }
        }
    }

    // Cleanup and return success
    outfile.close();
    return true;
}

Key Improvements & Explanations: #

1. File Handling: Added a filename parameter and uses std::ofstream to safely open/write to files, with error checking for failed file openings.
2. Input Validation: Ensures the clss vector length matches the data list length, and checks that row indices don't exceed the number of rows in each matrix.
3. Dense Matrix Handling: Directly extracts rows from NumericMatrix and writes elements with a comma delimiter (easily changeable to space/tab if needed).
4. Sparse Matrix Handling: Converts dgCMatrix to Armadillo's sp_mat, then converts the sparse row to a dense vector to preserve zero values (matching dense matrix output format).
5. Error Reporting: Uses Rcpp::Rcerr to print meaningful error messages to R's console, and cleans up the file handle before returning on failure.

Usage Example in R: #

library(Rcpp)
library(Matrix)

# Compile the function
sourceCpp("your_file_name.cpp")

# Create sample data
dense_mat <- matrix(rnorm(10), nrow = 5)
sparse_mat <- Matrix(rnorm(10), nrow = 5, sparse = TRUE)
data_list <- list(dense_mat, sparse_mat)
class_list <- sapply(data_list, class)

# Write rows to output.csv
write_rows(data_list, class_list, 5, "output.csv")

This will write rows in the order: row 0 of dense matrix → row 0 of sparse matrix → row 1 of dense matrix → row 1 of sparse matrix, etc.

内容的提问来源于stack exchange，提问作者Daniel Falbel