这个其实很多办法都可以做到，基本就是靠zcat后接awk，sed，或者wc命令，不过速度差强人意，比如这个

awk应该是这几个里最快的，但一个3个G左右的文件仍然运行了接近2分钟，考虑到还有几个30个G的文件需要测

磨刀不误砍柴工，我就写了个多线程的脚本，为了速度，就用了c++

#include <iostream>
#include <fstream>
#include <sstream>
#include <string>
#include <thread>
#include <vector>
#include <atomic>
#include <zlib.h>
#include <algorithm>
#include <sys/sysinfo.h>
#include <chrono>

std::atomic<unsigned long long> line_count(0);

void count_lines(const std::string& chunk, int chunk_size) {
    unsigned long long local_count = 0;
    for (int i = 0; i < chunk_size; ++i) {
        if (chunk[i] == '\n') {
            ++local_count;
        }
    }
    line_count += local_count;
}

int main(int argc, char* argv[]) {
    if (argc != 2) {
        std::cerr << "Usage: " << argv[0] << " <input_file>" << std::endl;
        return 1;
    }

    int max_threads = std::thread::hardware_concurrency() * 0.75;
    struct sysinfo sys_info;
    sysinfo(&sys_info);
    int BUFFER_SIZE = sys_info.freeram * 0.25 / max_threads;
    BUFFER_SIZE = (BUFFER_SIZE / max_threads) * max_threads;

    gzFile input_file = gzopen(argv[1], "rb");
    if (!input_file) {
        std::cerr << "Error opening file." << std::endl;
        return 1;
    }

    std::vector<std::thread> threads;
    std::string buffer(BUFFER_SIZE, 0);
    int bytes_read = 0;

    auto start_time = std::chrono::high_resolution_clock::now();

    while ((bytes_read = gzread(input_file, &buffer[0], BUFFER_SIZE)) > 0) {
        threads.push_back(std::thread(count_lines, buffer, bytes_read));

        if (threads.size() >= max_threads) {
            for (auto& t : threads) {
                t.join();
            }
            threads.clear();
        }
    }

    for (auto& t : threads) {
        t.join();
    }

    auto end_time = std::chrono::high_resolution_clock::now();
    auto duration = std::chrono::duration_cast<std::chrono::seconds>(end_time - start_time).count();

    gzclose(input_file);
    std::cout << "Number of lines: " << line_count << std::endl;
    std::cout << "Time taken: " << duration << " seconds" << std::endl;
    return 0;
}

这需要一个需要支持C++11或更高版本的编译器，比如我是gcc-v12.2，编译命令为

g++ -std=c++11 -o gz_rownum gz_rownum.cpp -lz -pthread

编译的脚本还有个好处，丢你的环境目录里，在哪都可以直接运行，执行一下看看效果

时间节省了一半，毕竟是大量i/o的工作，不可能太夸张，如果服务器参数好一点，效果应该会更明显