sqlite_orm性能测试简单记录

teaczjf

已于 2023-12-20 09:04:48 修改

阅读量1.3k

点赞数

分类专栏： linux 文章标签： sqlite 数据库 database

于 2022-03-16 10:30:57 首次发布

本文链接：https://blog.csdn.net/teaczjf/article/details/123520153

版权

linux 专栏收录该内容

7 篇文章 0 订阅

订阅专栏

参考资料
SQLite使用了某种锁定机制，可以防止您在数据库上执行并发操作。下面的连接介绍了sqlite锁的机制
https://blog.csdn.net/scyatcs/article/details/82690544

看下来，我感觉还是得上层自己加锁。 sqlite 并发处理

https://blog.csdn.net/cuit/article/details/21975807

https://blog.csdn.net/qingzhuyuxian/article/details/79767917?ops_request_misc=%257B%2522request%255Fid%2522%253A%2522164725837016780366572012%2522%252C%2522scm%2522%253A%252220140713.130102334…%2522%257D&request_id=164725837016780366572012&biz_id=0&utm_medium=distribute.pc_search_result.none-task-blog-2_allsobaiduend~default-1-79767917.142^v2pc_search_quality_down,143^v4control&utm_term=SQLite%E5%B9%B6%E5%8F%91&spm=1018.2226.3001.4187

性能测试

insert 插入

1.插入1000条数据，耗时6S

  TJ_GSCJ_DEVINFO.ID = 3;
    TJ_GSCJ_DEVINFO.DEV_NAME = "TJ_GSCJ";
    TJ_GSCJ_DEVINFO.DEV_CMD_TABLE_NAME = "TJ_GSCJ_CMDTABLE";
    TJ_GSCJ_DEVINFO.COMUNITATE_MODE = "TCP_SERVER";
    TJ_GSCJ_DEVINFO.TCP_IP = "127.0.0.1";
    TJ_GSCJ_DEVINFO.TCP_PORT = "5000";
    TJ_GSCJ_DEVINFO.UDP_IP_SRC = "Jameqs";
    TJ_GSCJ_DEVINFO.UDP_PORT_SRC = "Jameqs";
    TJ_GSCJ_DEVINFO.UDP_PORT_DST = "Jameqs";
    TJ_GSCJ_DEVINFO.UDP_IP_DST = "Jameqs";
    TJ_GSCJ_DEVINFO.UDPMC_IP = "Jameqs";
    TJ_GSCJ_DEVINFO.UDPMC_IP_RX = "Jameqs";
    TJ_GSCJ_DEVINFO.UDPMC_IP_TX = "Jameqs";
    TJ_GSCJ_DEVINFO.UDPMC_PORT_RX = "Jameqs";
    TJ_GSCJ_DEVINFO.UDPMC_PORT_TX = "Jameqs";


printf_init_log("start  sqlite_orm_write_test");
    int key = 1000; //写1000条大概，耗时6
    while (key != 0)
    {
        storage.insert(TJ_GSCJ_DEVINFO); //大概在4-8ms
        key--;
    }
    printf_init_log("#############################start  read sqlite_orm_write_test");

2.单条数据插入时间

  storage.replace(SQLORM_XPRO_DEVINFO{1, "James", "James", "Houston", "Houston", "Houston", "Houston", "Houston", "Houston", "Houston", "Houston", "Houston", "Houston", "Houston", "Houston"}); //大概在6-8ms

  storage.insert(TJ_GSCJ_DEVINFO);  //大概在4-8ms

读取

1.读取耗时

同一时间只有一个线程在做读取

        //  读取次数     耗时
        //  90000        57-58ms
        // 一次读6000    4-5ms
        // 1000          2m
        // 500           1m
printf_init_log("#############################start  read sqlite_orm_write_test");
auto simpleRows = storage.select(columns(&SQLORM_XPRO_DEVINFO::ID, &SQLORM_XPRO_DEVINFO::DEV_NAME, &SQLORM_XPRO_DEVINFO::DEV_CMD_TABLE_NAME)); //几乎不耗时 小于1ms
printf_init_log("#############################end read  sqlite_orm_write_test");

2.一对一的读取

循环读取,不带延时。读取正常，没有闪退

void *thread_GSCJQ1(void *arg)
{
    while (1)
    {
        LOG(INFO) << "#############################start  read sqlite_orm_write_test";
        try
        {
            auto simpleRows = storage1.select(columns(&SQLORM_XPRO_DEVINFO::ID, &SQLORM_XPRO_DEVINFO::DEV_NAME, &SQLORM_XPRO_DEVINFO::DEV_CMD_TABLE_NAME)); //几乎不耗时 小于1ms
        }
        catch (const std::system_error &e)
        {
            cout << "11111111111" << e.what() << endl;
        }
    }
}

3.多对一的读取

起两个线程，循环读取,不带延时。读取正常，没有闪退

注意，两个线程的句柄需要不同。

void *thread_GSCJQ1(void *arg)
{
    while (1)
    {
        LOG(INFO) << "#############################start  read sqlite_orm_write_test";
        try
        {
            auto simpleRows = storage1.select(columns(&SQLORM_XPRO_DEVINFO::ID, &SQLORM_XPRO_DEVINFO::DEV_NAME, &SQLORM_XPRO_DEVINFO::DEV_CMD_TABLE_NAME)); //几乎不耗时 小于1ms
        }
        catch (const std::system_error &e)
        {
            cout << "11111111111" << e.what() << endl;
        }
    }
}


void *thread_GSCJQ2(void *arg)
{
    while (1)
    {
        LOG(TRACE) << "__________________start  read sqlite_orm_write_test";
        try
        {
            auto simpleRows = storage2.select(columns(&SQLORM_XPRO_DEVINFO::ID, &SQLORM_XPRO_DEVINFO::DEV_NAME, &SQLORM_XPRO_DEVINFO::DEV_CMD_TABLE_NAME)); //几乎不耗时 小于1ms
        }
        catch (const std::system_error &e)
        {
            cout << "22222222" << e.what() << endl;
        }
    }
}