稀疏矩阵的快速转置c语言代码

以下是稀疏矩阵快速转置的C语言代码实现：

#include <stdio.h>
#include <stdlib.h>

void sparse_matrix_transpose(int *rows, int *cols, int *vals, int n, int m, int nz)
{
    int i, j, k, col_count, *col_counts, *col_starts, *new_rows, *new_cols, *new_vals;

    // 统计每一列的非零元素个数
    col_counts = (int*)calloc(m, sizeof(int));
    for (i = 0; i < nz; i++) {
        col_counts[cols[i]]++;
    }

    // 计算每一列的起始位置
    col_starts = (int*)calloc(m + 1, sizeof(int));
    col_starts[0] = 0;
    for (i = 1; i <= m; i++) {
        col_starts[i] = col_starts[i - 1] + col_counts[i - 1];
    }

    // 创建新的稀疏矩阵
    new_rows = (int*)calloc(nz, sizeof(int));
    new_cols = (int*)calloc(nz, sizeof(int));
    new_vals = (int*)calloc(nz, sizeof(int));
    for (i = 0; i < n; i++) {
        for (j = rows[i]; j < rows[i + 1]; j++) {
            k = col_starts[cols[j]];
            new_rows[k] = cols[j];
            new_cols[k] = rows[i];
            new_vals[k] = vals[j];
            col_starts[cols[j]]++;
        }
    }

    // 将新的稀疏矩阵复制回原数组
    for (i = 0; i < nz; i++) {
        rows[i] = new_rows[i];
        cols[i] = new_cols[i];
        vals[i] = new_vals[i];
    }

    // 释放内存
    free(col_counts);
    free(col_starts);
    free(new_rows);
    free(new_cols);
    free(new_vals);
}

该代码实现的思路与Python代码类似，先统计每一列的非零元素个数，然后计算每一列的起始位置，最后创建新的稀疏矩阵。需要注意的是，C语言中需要手动申请和释放内存。