原文： C#中DataTable中的Compute方法使用收集
Compute函数的参数就两个：Expression，和Filter。
Expresstion是计算表达式，关于Expression的详细内容请看这里“http://msdn2.microsoft.com/zh-cn/library/system.data.datacolumn.expression(VS.80).aspx”。而Filter则是条件过滤器，类似sql的Where条件。
DataTable dt = new DataTable();
//嵌套的三元运算 牛叉到五体投地
object obj = dt.Compute("iif(1000=5,1000,iif(100>100,4001,2000))", null);
Response.Write(obj);
System.Data.DataTable table = new DataTable();
//计算常量，可以没有初始化列
object test = table.Compute("1+1", "");
Console.WriteLine(test);
string a = "123";
System.Double b = 123;
decimal c = 123m;
Console.WriteLine(Convert.ToDecimal(a));
//test=2;
test = table.Compute("1+1", "false");
Console.WriteLine(test);
//test=2;常数计算和filter无关
test = table.Compute("abs(1)", "");
Console.WriteLine(test);
//test=null，不知道为这个什么没有报错，而且返回null,其他的数学函数都会抱错
test = table.Compute("2%2", "");
Console.WriteLine(test);
//test=0;
//其他函数参考下面的计算列
//初始化datatale
table.Columns.Add("id", typeof(string));
table.Columns.Add("value", typeof(int));
for (int i = 1; i <= 10; i++)
{
System.Data.DataRow dRow = table.NewRow();
dRow["id"] = "id" + i.ToString();
dRow["value"] = i;
table.Rows.Add(dRow);
}
//test = table.Compute("value+1", "true");
/**/
////抛出异常，这里必须是聚合函数
//*************************************支持的聚合函数**********************//
//求数量
test = table.Compute("count(id)", "false");
Console.WriteLine(test);
//test=0;
test = table.Compute("count(id)", "true");
Console.WriteLine(test);
//test=10;
//求和
test = table.Compute("sum(value)", "");
Console.WriteLine(test);
//test=55;
//test = table.Compute("sum(id)","");
/**/
////抛出异常，这里不能是string
//平均
test = table.Compute("avg(value)", "");
Console.WriteLine(test);
//test=5;
//最小
test = table.Compute("min(value)", "");
Console.WriteLine(test);
//test=1;
//最大
test = table.Compute("max(value)", "");
Console.WriteLine(test);
//test=10;
//统计标准偏差
test = table.Compute("StDev(value)", "");
Console.WriteLine(test);
//test=3.02765035409749
//统计方差
test = table.Compute("Var(value)", "");
Console.WriteLine(test);
//test=9.16666666666667
//复杂计算
test = table.Compute("max(value)/sum(value)", "");
Console.WriteLine(test);
//test=0.181818181818182
/**/
/*******************************************计算列*************************/
System.Data.DataColumn column = new DataColumn("exp1", typeof(float));
table.Columns.Add(column);
//简单计算
column.Expression = "value*2";
test = table.Select("id='id1'")[0]["exp1"];
Console.WriteLine(test);
//test=2;
//字符串函数
column.Expression = "len(id)";
test = table.Select("id='id1'")[0]["exp1"];
Console.WriteLine(test);
//test=3;
//字符串函数
column.Expression = "len(' '+id+' ')";
test = table.Select("id='id1'")[0]["exp1"];
Console.WriteLine(test);
//test=5;
//字符串函数
column.Expression = "len(trim(' '+id+' '))";
test = table.Select("id='id1'")[0]["exp1"];
Console.WriteLine(test);
//test=3;
//字符串函数
column.Expression = "substring(id,3,len(id)-2)";
test = table.Select("id='id1'")[0]["exp1"];
Console.WriteLine(test);
//test=1; //substring的起始字符位置为1不是0
//类型转换
column.Expression = "convert(substring(id,3,len(id)-2),'System.Int32')*1.6";
test = table.Select("id='id1'")[0]["exp1"];
Console.WriteLine(test);
//test=1.6;
//相当于sqlserver的isnull
column.Expression = "isnull(value,10)";
test = table.Select("id='id1'")[0]["exp1"];
Console.WriteLine(test);
//test=1;
//三元运算符,相当于sqlserver的case when
column.Expression = "iif(value>5,1000,2000)";
test = table.Select("id='id1'")[0]["exp1"];
Console.WriteLine(test);
//test=2000;
//like运算符
column.Expression = "iif(id like '%1',1000,2000)";
test = table.Select("id='id1'")[0]["exp1"];
Console.WriteLine(test);
//test=1000;
//in运算符
column.Expression = "iif(id not in('id1'),1000,2000)";
test = table.Select("id='id1'")[0]["exp1"];
Console.WriteLine(test);
//test=2000;
//嵌套的三元运算
column.Expression = "iif(value>5,1000,iif(id like '%1',4000,2000))";
test = table.Select("id='id1'")[0]["exp1"];
Console.WriteLine(test);
//test=4000;
//客户端计算所占总数的百分比
column.Expression = "value/sum(value)";
test = table.Select("id='id1'")[0]["exp1"];
Console.WriteLine(test);
//test=0.01818182
//客户端计算差值,比如nba常规赛的胜场差
column.Expression = "max(value)-value";
test = table.Select("id='id1'")[0]["exp1"];
Console.WriteLine(test);
//test=9
//***********************父子表计算*************************************/
//初始化子表,父子表关系
DataTable tableChild = new DataTable();
tableChild.Columns.Add("id", typeof(string));
tableChild.Columns.Add("value", typeof(int));
System.Data.DataSet ds = new DataSet();
ds.Tables.Add(tableChild);
ds.Tables.Add(table);
DataRelation relation = new DataRelation("relation", table.Columns["id"], tableChild.Columns["id"]);
ds.Relations.Add(relation);
for (int i = 1; i <= 10; i++)
{
System.Data.DataRow dRow = tableChild.NewRow();
dRow["id"] = "id1";
dRow["value"] = i;
tableChild.Rows.Add(dRow);
}
//计算子表记录数
column.Expression = "count(child(relation).value)";
test = table.Select("id='id1'")[0]["exp1"];
Console.WriteLine(test);
//test=10;
//计算父子表的百分比
column.Expression = "value/sum(child(relation).value)";
test = table.Select("id='id1'")[0]["exp1"];
Console.WriteLine(test);
//test=0.01818182;
//计算父子表的差值,比如父表为库存数量，子表为订购数量，计算得出需要补充的数量
column.Expression = "iif(value-sum(child(relation).value)>0,0,value-sum(child(relation).value))";
test = table.Select("id='id1'")[0]["exp1"];
Console.WriteLine(test);
//test=-54;
//比较遗憾的是没有发现能够计算同比和环比的方法，而且计算列无法作为约束
//结束，DataTable可以让你尽量发挥聪明才智来减少繁杂的sql语句并且减轻服务器计算符合