分组聚合

nicye大约 4 分钟约 1194 字

分组聚合

IFreeSql fsql; //如何创建请移步入门文档

class Topic 
{
    [Column(IsIdentity = true, IsPrimary = true)]
    public int Id { get; set; }
    public int Clicks { get; set; }
    public string Title { get; set; }
    public DateTime CreateTime { get; set; }
}

1、单表分组

var list = fsql.Select<Topic>()
    .GroupBy(a => new { tt2 = a.Title.Substring(0, 2), mod4 = a.Id % 4 })
    .Having(g => g.Count() > 0 && g.Avg(g.Key.mod4) > 0 && g.Max(g.Key.mod4) > 0)
    .Having(g => g.Count() < 300 || g.Avg(g.Key.mod4) < 100)
    .OrderBy(g => g.Key.tt2)
    .OrderByDescending(g => g.Count())
    .ToList(g => new 
    {
        g.Key, 
        cou1 = g.Count(), 
        arg1 = g.Avg(g.Value.Clicks), 
        arg2 = g.Sum(g.Value.Clicks > 100 ? 1 : 0)
    });
//SELECT 
//substr(a.`Title`, 1, 2) as1,
//(a.`Id` % 4) as2, 
//count(1) as3, 
//avg(a.`Clicks`) as4, 
//sum(case when a.`Clicks` > 100 then 1 else 0 end) as5 
//FROM `Topic` a 
//GROUP BY substr(a.`Title`, 1, 2), (a.`Id` % 4) 
//HAVING (count(1) > 0 AND avg((a.`Id` % 4)) > 0 AND max((a.`Id` % 4)) > 0) AND (count(1) < 300 OR avg((a.`Id` % 4)) < 100)
//ORDER BY substr(a.`Title`, 1, 2), count(1) DESC

不分组求聚合值，请使用 ToAggregate 替代 ToList

var list = fsql.Select<Topic>()
    .ToAggregate(g => new 
    {
        cou1 = g.Count(), 
        arg1 = g.Avg(g.Key.Clicks), 
        arg2 = g.Sum(g.Key.Clicks > 100 ? 1 : 0)
    });

2、多表分组

var list = fsql.Select<Topic, Category, Area>()
    .InnerJoin((a, b, c) => b.Id == a.CategoryId)
    .InnerJoin((a, b, c) => c.Id == b.AreaId)
    .GroupBy((a, b, c) => new { a.Title, c.Name })
    .Having(g => g.Count() < 300 || g.Avg(g.Value.Item1.Clicks) < 100)
    .ToList(g => new { count = g.Count(), Name = g.Key.Name });
//SELECT count(1), c.name
//FROM topic a
//LEFT JOIN cagetory b ON b.id = a.category_id
//LEFT JOIN area c ON c.id = b.area_id
//GROUP BY a.title, c.name
//HAVING count(1) < 300 AND avg(a.clicks) < 100

g.Value.Item1 对应 Topic
g.Value.Item2 对应 Category
g.Value.Item3 对应 Area

说明	方法	SQL
总数	.Count()	select count(*) from ...
求和	.Sum(a => a.Score)	select sum([Score]) from ...
平均	.Avg(a => a.Score)	select avg([Score]) from ...
最大值	.Max(a => a.Score)	select max([Score]) from ...
最小值	.Min(a => a.Score)	select min([Score]) from ...

lambda	sql	说明
SqlExt.IsNull(id, 0)	isnull/ifnull/coalesce/nvl	兼容各大数据库
SqlExt.DistinctCount(id)	count(distinct id)
SqlExt.GreaterThan	>	大于
SqlExt.GreaterThanOrEqual	>=	大于或等于
SqlExt.LessThan	<	小于
SqlExt.LessThanOrEqual	<=	小于
SqlExt.EqualIsNull	IS NULL	是否为 NULL
SqlExt.Case(字典)	case when .. end	根据字典 case
SqlExt.GroupConcat	group_concat(distinct .. order by .. separator ..)	MySql
SqlExt.FindInSet	find_in_set(str, strlist)	MySql
SqlExt.StringAgg	string_agg(.., ..)	PostgreSQL
SqlExt.Rank().Over().PartitionBy().ToValue()	rank() over(partition by xx)	开窗函数
SqlExt.DenseRank().Over().PartitionBy().ToValue()	dense_rank() over(partition by xx)
SqlExt.Count(id).Over().PartitionBy().ToValue()	count(id) over(partition by xx)
SqlExt.Sum(id).Over().PartitionBy().ToValue()	sum(id) over(partition by xx)
SqlExt.Avg(id).Over().PartitionBy().ToValue()	avg(id) over(partition by xx)
SqlExt.Max(id).Over().PartitionBy().ToValue()	max(id) over(partition by xx)
SqlExt.Min(id).Over().PartitionBy().ToValue()	min(id) over(partition by xx)
SqlExt.RowNumber(id).Over().PartitionBy().ToValue()	row_number(id) over(partition by xx)

3、分组第一条记录

fsql.Select<User1>()
    .Where(a => a.Id < 1000)
    .WithTempQuery(a => new
    {
        item = a,
        rownum = SqlExt.RowNumber().Over().PartitionBy(a.Nickname).OrderBy(a.Id).ToValue()
    })
    .Where(a => a.rownum == 1)
    .ToList();

提示：支持多表嵌套查询，fsql.Select<User1, UserGroup1>()

SELECT *
FROM (
    SELECT a.[Id], a.[Nickname], row_number() over( partition by a.[Nickname] order by a.[Id]) [rownum]
    FROM [User1] a
    WHERE a.[Id] < 1000
) a
WHERE (a.[rownum] = 1)

如果数据库不支持开窗函数，可以使用分组嵌套查询解决：

fsql.Select<User1>()
    .Where(a => a.Id < 1000)
    .GroupBy(a => a.Nickname)
    .WithTempQuery(g => new { min = g.Min(g.Value.Id) })
    .From<User1>()
    .InnerJoin((a, b) => a.min == b.Id)
    .ToList((a, b) => b);

SELECT b.[Id], b.[Nickname] 
FROM ( 
    SELECT min(a.[Id]) [min] 
    FROM [User1] a 
    WHERE a.[Id] < 1000 
    GROUP BY a.[Nickname] ) a 
INNER JOIN [User1] b ON a.[min] = b.[Id]

查看更多《嵌套查询》文档

4、Aggregate

Distinct

var list = fsql.Select<Topic>()
    .Aggregate(a => Convert.ToInt32("count(distinct title)"), out var count)
    .ToList();

SELECT count(distinct title) as1 FROM "Topic" a

SELECT a."Id", a."Clicks", a."Title", a."CreateTime" FROM "Topic" a

SqlExt.DistinctCount

fsql.Select<Topic>()
    .Aggregate(a => SqlExt.DistinctCount(a.Key.Title), out var count);

SELECT count(distinct a."title") as1 FROM "Topic" a

ToAggregate + SqlExt.DistinctCount

var distinctAggregate = fsql.Select<Topic>().ToAggregate(a => new
    {
        TitleCount = SqlExt.DistinctCount(a.Key.Title),
        ClicksCount= SqlExt.DistinctCount(a.Key.Clicks),
    }
);

SELECT count(distinct a."Title") as1, count(distinct a."Clicks") as2 FROM "Topic" a

5、导航属性分组

假如 Topic 有导航属性 Category，Category 又有导航属性 Area，导航属性分组代码如下：

var list = fsql.Select<Topic>()
    .GroupBy(a => new { a.Clicks, a.Category })
    .ToList(g => new { g.Key.Category.Area.Name });

注意：如上这样编写，会报错无法解析 a.Key.Category.Area.Name，解决办法使用 Include：

var list = fsql.Select<Topic>()
    .Include(a => a.Category.Area)
    //必须添加此行，否则只分组 Category 而不包含它的下级导航属性 Area

    .GroupBy(a => new { a.Clicks, a.Category })
    .ToList(g => new { g.Key.Category.Area.Name });

但是，你还可以这样解决：

var list = fsql.Select<Topic>()
    .GroupBy(a => new { a.Clicks, a.Category, a.Category.Area })
    .ToList(g => new { g.Key.Area.Name });

API

方法	返回值	参数	描述
ToSql	string		返回即将执行的 SQL 语句
ToList<T>	List<T>	Lambda	执行 SQL 查询，返回指定字段的记录，记录不存在时返回 Count 为 0 的列表
ToList<T>	List<T>	string field	执行 SQL 查询，返回 field 指定字段的记录，并以元组或基础类型(int,string,long)接收，记录不存在时返回 Count 为 0 的列表
ToAggregate<T>	List<T>	Lambda	执行 SQL 查询，返回指定字段的聚合结果（适合不需要 GroupBy 的场景）
Sum	T	Lambda	指定一个列求和
Min	T	Lambda	指定一个列求最小值
Max	T	Lambda	指定一个列求最大值
Avg	T	Lambda	指定一个列求平均值
【分组】
GroupBy	<this>	Lambda	按选择的列分组，GroupBy(a => a.Name)
GroupBy	<this>	string, parms	按原生sql语法分组，GroupBy("concat(name, @cc)", new { cc = 1 })
Having	<this>	string, parms	按原生sql语法聚合条件过滤，Having("count(name) = @cc", new { cc = 1 })
【成员】
Key			返回 GroupBy 选择的对象
Value			返回主表或 From<T2,T3....> 的字段选择器

分组聚合

# 分组聚合

# 1、单表分组

# 2、多表分组

# 3、分组第一条记录

# 4、Aggregate

# Distinct

# SqlExt.DistinctCount

# ToAggregate + SqlExt.DistinctCount

# 5、导航属性分组

# API

分组聚合

1、单表分组

2、多表分组

3、分组第一条记录

4、Aggregate

Distinct

SqlExt.DistinctCount

ToAggregate + SqlExt.DistinctCount

5、导航属性分组

API