问题描述
寻找有关使用Thread的建议。
我有控制器，该控制器在DB中进行查询并转换List中的选定数据。但是在某些情况下，我需要几行的额外信息。我不想为每行循环查询。 如果我要启动新线程，它将与我的主循环并行进行额外的查询，并将额外的信息添加到列表中。好还是不好。或者最好先获取所有数据。
示例：
[HttpPost]
public async Task<IActionResult> ReestrRazr(DateTime date)
{
List<SomeClass> arr = new List<SomeClass>();
string query = " SELECT [table].[id],[table].[snumber],[table].[datein],[table].[datefileeate],[table].[istatusU],[table].[datezamechania],[table].[type],[table].[isubtypeU],";
query += " [tableT3040].[content] as [address],[tableT3040].[nazvanie] as [objNazvanie],";
query += " [tableT101].[nazvanie] as zastrName,";
query += " [adress].[rajon],";
query += " [objT3].[idcomlinknypredstavitelpurpose] as [nadzor] ";
query += " FROM ((((((((([table] ";
query += " JOIN [link] on (([table].[id]=[link].[idlinkper]) and ([link].[ideletestate]=0)) ) ";
query += " JOIN [file] on (([file].[id]=[link].[idlink]) and ([file].[sfileextension] like ('%pdf%')) and ([file].[ideletestate]=0)) ) ";
query += " JOIN [obj] ON ([table].[identity]=[obj].[id] AND [obj].[ideletestate]=0) ) ";
query += " LEFT JOIN [objT] as [objT3] on (([objT3].[identity]=[obj].[id]) and ([objT3].[type]=3) and ([objT3].[ideletestate]=0)) )";
query += " LEFT JOIN [adress] ON (([obj].[idaddress]=[adress].[id]) and ([adress].[ideletestate]=0) and ([adress].[workstate]=1)) ) ";
query += " LEFT JOIN [objT] as [objT119] on (([objT119].[identity]=[obj].[id]) and ([objT119].[type]=119) and ([objT119].[ideletestate]=0) and ([objT119].[idcomlinknypredstavitelpurpose] = 1)) )";
query += " LEFT JOIN [tableT] as [tableT3040] on (([table].[id]=[tableT3040].[idlinkper]) and ([tableT3040].[type]=3040) and ([tableT3040].[ideletestate]=0)) ) ";
query += " LEFT JOIN [tableT] as [tableT101] on (([table].[id]=[tableT101].[idlinkper]) and ([tableT101].[type]=101) and ([tableT101].[nazvanie] <> '') and ([tableT101].[nazvanie] is not null) and ([tableT101].[ideletestate]=0)) )";
query += " left join [ae] on (([table].[idcomlinkny]=[ae].[id]) and ([ae].[ideletestate]=0)) ) ";
DataTable dt = Execute4Table(query + $" where date >'{date.tostring("yyyyMMdd")}'")
foreach (DaTarow row in dt.Rows)
{
int id = Convert.ToInteger(row[0]);
int type = Convert.ToInteger(row[1]);
string number= Convert.ToString(row[2]);
SomeClass sc = new SomeClass();
sc.id=id;
sc.number = number;
if (type==50)
{
number = Convert.ToString(ExecuteScalar("select number from table WHERE idparent="+ id.tostring()));
}
arr.Add(sc);
}
}
谢谢！（对不起我的英语） 附言实际查询有8个JOIN，获得所有必要的信息将导致JOIN的数量增加到12-13。这对性能不利。
解决方法
您将无法从同一DbContext（请求范围）运行并行查询。我可以对一个查询执行left outer join，它将为您带来所需的所有数据，其他行为null，或者允许客户端对这些行（请求/行）提出额外的请求。尽管我会坚持第一个想法，但前提是不会发现那些“额外”会导致查询的基本版本大大降低速度。 3ed选项是仅对主查询使用await，从内存结果中收集选定的ID，然后对IN进行第二次查询，以稍后在内存中分配值。