sqlite 版本更新维护, 表结构判断, 更新

sqlite会自动维护一个系统表sqlite_master，该表存储了我们所创建的各个table, view, trigger等等信息。

sqlite_master表数据字段：
type:  类型，取值一般为table, view
name:
tbl_name:   表名
rootpage：

sql：创建表或者视图的sql语句，可以从该sql语句中判断某字段是否存在

sqlite_master表结构如下：

    CREATE TABLE sqlite_master (
    type TEXT,
    name TEXT,
    tbl_name TEXT,
    rootpage INTEGER,
    sql TEXT
    )
————————————————
版权声明：本文为CSDN博主「小向光」的原创文章，遵循 CC 4.0 BY-SA 版权协议，转载请附上原文出处链接及本声明。
原文链接：https://blog.csdn.net/wuyou1336/article/details/53770799

---

判断数据库是否存在, 表格式是否正确(版本更新时,很重要)

数据库更新,可以增加table_version 来进行判断

table_version
(
	tb_name  varchar(32)
	tb_version UNSIGNED INTEGER NOT NULL
	comment varchar(128)
)

保存每个table 版本日期,  有不一致,就修改,  历史版本在代码中有记录,可以按当前记录,去修改表结构.

(如果差别太多, 就直接备份旧表,人工导出csv数据, 然后 创建新表,再导入数据)重新写代码修改 并导入,或者直接改csv 文件

 sqlite_master中得到 表结构的SQL语句，自己再逐个分析  

//表名已经确认
//解析括号,
//, 得到单个表结构数据,  逐个比较就行

或者直接全部比较?

CREATE TABLE [Table_1004A] (
                        [ChipID] UNSIGNED INTEGER NOT NULL PRIMARY KEY,
                        [ChipType] VARCHAR(32) NOT NULL,
                        [BatchID] UNSIGNED INTEGER NOT NULL,

                        [T1_TestMode] VARCHAR(32) ,
                        [T1_Timestamp] UNSIGNED INTEGER,
                        [T1_Temp] SMALLINT,
                        [T1_ErrSaAddrNum] UNSIGNED INTEGER,
                        [T1_ErrSaAddr] VARCHAR(256),
                        [T1_ChipResult] VARCHAR(64),
                        [T1_HandlerResult] UNSIGNED TINYINT,

                        [T2_TestMode] VARCHAR(32),
                        [T2_Timestamp] UNSIGNED INTEGER,
                        [T2_Temp] SMALLINT,
                        [T2_ErrSaAddrNum] UNSIGNED INTEGER,
                        [T2_ErrSaAddr] VARCHAR(256),
                        [T2_ChipResult] VARCHAR(64),
                        [T2_RepairedSaAddr] VARCHAR(256),
                        [T2_HandlerResult] UNSIGNED TINYINT ,

                        [Comment] VARCHAR(256)
                        )

SELECT name FROM sqlite_master
WHERE type=’table’
ORDER BY name; 

 //ok
select * from sqlite_master;
select type,name,sql from sqlite_master; 

select type,name,sql from sqlite_master order by name; 

select type,name,sql from sqlite_master WHERE name='Table_1004A';

===================

//默认升序
SELECT NumID,ChipID,BatchID,T1_Timestamp FROM main.Table_1004A WHERE Table_1004A.NumID >= 1000 ORDER BY ChipID;

1010	1301000100001000	100	1570877172
1011	1301000100002001	100	1570877359
1016	1301000100004001	100	1570879295
1015	1301000100004002	100	1570879275
1014	1301000100004004	100	1570879275
1013	1301000100004010	100	1570879275
1012	1301000100004020	100	1570879274

/*==========*/

//降序
SELECT NumID,ChipID,BatchID,T1_Timestamp FROM main.Table_1004A WHERE Table_1004A.NumID >= 1000 ORDER BY ChipID DESC;

1012	1301000100004020	100	1570879274
1013	1301000100004010	100	1570879275
1014	1301000100004004	100	1570879275
1015	1301000100004002	100	1570879275
1016	1301000100004001	100	1570879295
1011	1301000100002001	100	1570877359
1010	1301000100001000	100	1570877172

insert 硬盘的IO速度,最受影响. (如果primier key 是chipID, chipID插入顺序不规律, 但是DB内部会使chipID变有序)
insert 的如果chipID 重复,就crash. 所以还要先查询下,chipID是否已经有了.(T2 也先查询下,没有也报错)

查询时间因为chipID 用RBTree建立了有序索引,所以效率还算可以, 5W记录 8ms.

如果按照之前的表设计,  numID 为 primier key autoincrease, (用NumID 为索引建立RBTree)则insert 变快(因为每次都是自增的值,不用改位置), query变慢 chipID无序,

//HD(7200)   写入速度:11个record每秒
SELECT COUNT(*) FROM "Table_1004A" WHERE T1_Timestamp =1571039500;

//SSD(TLC)   写入速度:50个record每秒
SELECT COUNT(*) FROM "Table_1004A" WHERE T1_Timestamp =1571040594;

//Memory   写入速度:大约14757个record每秒(100W/70s),cpu i5-8400@2.8G 负载25%(可能只用到了单核心,多核心可以加倍), 内存占用100MB左右
SELECT COUNT(*) FROM "Table_1004A" WHERE T1_Timestamp =1571122590;  //查询时间118ms

SELECT * FROM "Table_1004A" WHERE ChipID=1301000100001000;

//单值修改 

UPDATE employees
SET lastname = 'Smith'
WHERE employeeid = 3;

//多值修改
UPDATE table
SET column_1 = new_value_1,
    column_2 = new_value_2
WHERE
    search_condition
ORDER column_or_expression
LIMIT row_count OFFSET offset;

/*==========*/
UPDATE Table_1004A SET T2_Timestamp=1571041417 WHERE ChipID=1301000100001011;

SELECT * FROM "Table_1004A" WHERE T2_Timestamp NOTNULL;

SELECT * FROM "Table_1004A" WHERE T2_Timestamp=1571041417;

Paginate 页码  共多少页,每页多少数据
https://sqlkata.com/docs/execution/fetching-records

var page1 = query.Paginate(1);

foreach(var item in page1.Each)
{
    // print items in the first page
}

var page2 = page1.Next(); // same as query.Paginate(2)

foreach(var item in page2.Each)
{
    // print items in the 2nd page
}

var users = query.Paginate(1, 10);

foreach(var user in users.Each)
{
    Console.WriteLine($"Id: {user.Id}, Name: {user.Name}");
}

//找到最大的rowid(如果是自定义的primaryKey, 比如 NumID  则都可以用,  反正是alias 别名)
SELECT  *,MAX(rowid) FROM Table_1004A;//推荐
SELECT  *,MAX(NumID  ) FROM Table_1004A;//推荐

SELECT rowid FROM Table_1004A ORDER BY ROWID DESC LIMIT(1);
SELECT *,rowid FROM Table_1004A ORDER BY ROWID DESC LIMIT(1);

//SELECT *,rowid  FROM Table_1004A LIMIT 53000,1000

//返回最后一次insert的rowid,如果没有插入就返回0 (开机重启后也返回0, 是记录在内存中)
SELECT LAST_INSERT_ROWID();

//SELECT *  FROM Table_1004A where rowid=LAST_INSERT_ROWID();

SELECT *,rowid "NAVICAT_ROWID" FROM "main"."Table_1004A" LIMIT 53000,1000

//===============

//时间: 0.03s
INSERT INTO "main"."Table_1004A"("ChipID", "ChipType", "BatchID", "T1_TestMode", "T1_Timestamp", "T1_Temp", "T1_ErrSaAddrNum", "T1_ErrSaAddr", "T1_ChipResult", "T1_HandlerResult",
"T2_TestMode", "T2_Timestamp", "T2_Temp", "T2_ErrSaAddrNum", "T2_ErrSaAddr", "T2_ChipResult", "T2_RepairedSaAddr", "T2_HandlerResult", "Comment")
VALUES (1301000100005001, 'XM1004_A_DWL', 100, 'ArrayWR', 1571106878, 120, 5, '1,2,3,4,5', 'Finish_MarchC', 1, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL);

//INSERT后,  LAST_INSERT_ROWID() 结果显示53102 + 显示最后插入信息
	1. SELECT LAST_INSERT_ROWID();
	2. SELECT *,rowid FROM Table_1004A WHERE rowid= 53102;

//2句话,一句搞定
SELECT *,rowid FROM Table_1004A WHERE rowid=LAST_INSERT_ROWID();

            //SQLiteConfigDbOpsEnum sQLiteConfigDbOpsEnum = SQLiteConfigDbOpsEnum.SQLITE_DBCONFIG_NONE;
            //connection.SetConfigurationOption(sQLiteConfigDbOpsEnum, );

            //var users = new XQuery(connection, compiler).From("Users").Limit(10).Get();

            var db = new QueryFactory(connection, compiler);
            // You can register the QueryFactory in the IoC container
            var query = db.Query("UsersTable").Where("Id", 1).Where("Status", "Active").First();

            /*
            var query = new Query()
                    .Select("client.Id")
                    .From("schema.client as c")
                    .Join("schema.order as o", "c.Id", "o.ClientId")
                    .WhereIn("c.Id", new[] { 1, 2, 3 });
            */