如何在LINQ中加快查询二进制搜索

Question

I use Linq select statement to get file as byte from the database. It take long time to get the byte array from my Files Table.

I'm using this code:

Fun<Files,bool> Filter  = delegate(Files x)
{
     return x.FileID == 10;
}; 
Files File = DAL.FilesDAL.GetFile(Filter).First();


public static List<Files> GetFile(Func<Files,bool> lambda){
return db.Where(lamnda).ToList();
}

For 1M file size it take up to 1m. That too long for my clients.

How I can improve the speed of that query?

Answer 1

Looks like you're missing the fact, that using Func<> makes your query be executed as LINQ to Objects. That means entire table is fetched into application memory and then filtering is performed by the application, not by the database. To make the filter execute at database it should be passed as Expression<Func<>> instead:

public static List<Files> GetFile(Expression<Func<Files,bool>> lambda){
    return db.Where(lamnda).ToList();
}

I assumed, that db here is IQueryable<Files> .

To call it use:

Files File = DAL.FilesDAL.GetFile(x => x.Filter == 10).First();

To make it even more descriptive, you should probably change your method to only return one Files item. Method name is GetFile , so I would expect it to return one file, not collection of files:

public static Files GetFile(Expression<Func<Files,bool>> lambda){
    return db.FirstOrDefault(lamnda);
}

usage:

var File = DAL.FilesDAL.GetFile(x => x.Filter == 10);

Or to increase semantics, you can refactor the method to be GetFileById and take Id , not an expression:

public static Files GetFileById(int Id)
{
    return db.FirstOrDefault(x => x.FileId == id);
}

usage

var File = DAL.FilesDAL.GetFileById(10);

Answer 2

How I can improve the speed of that query?

Faster server. Faster network. Do not get the binary content you do not need.

SImple like that. NOthing you can do on LINQ level that will magically fix a bad design.

Generally:

Btw., a bad mistake - you return ToList in GetFiles. Never do that. Either let it a queryable, so the user can add conditions and order and grouping or return an enumerable (which is way less powerfull) but ToList means all gets into memory - even if the user then does not want a list.

And man, the way to hand in the lambdaa is just redundant and limiting. Bad design.