I'm using Entity Framework core 6 with SQL Server. I have the following query:
dbContext.Table1
.Where(t1 => some_condition(t1.some_property))
.Select(t1 => new
{
Field1 = t1.some_property,
Field2 = dbContext.Table2.Where(t2 => some_condition(t1.some_property, t2.some_property)).First(), // <-- this is my question
});
The thing is that this subquery dbContext.Table2.Where(t2 => some_condition(t1, t2)).First(), is used in many other places. I need to extract it somehow and call it instead of copy-pasting it everywhere.
I tried to extract it in a function like this:
public string Get(string some_property) => context.Table2
.Where(t => some_condition(some_property, t.some_property))
.First();
dbContext.Table1
.Where(t1 => some_condition(t1.some_property))
.Select(t1 => new
{
Field1 = t1.some_property,
Field2 = Get(t1.some_property)
});
But with that I get the subquery to execute separately and for every element in the collection (N+1 problem. I was expecting it though).
Is there a way to achieve that ? maybe using expression trees (which I'm not yet very familiar with).
Related
I am trying to build a query using asp.net core c#
https://www.reflectionit.nl/blog/2017/paging-in-asp-net-core-mvc-and-entityframework-core
I trying to do a filtering however I need the data from another table which have my unique id
var result = _context.UserRoles.Where(y => y.RoleId.Contains(selectedRoles.Id)); // Retrieve the the userid i have from another table with the selected roleid
var query = _context.Users.Where(x => //I have already tried contains, where join );
If there is a site where i can learn this query please recommend. "Join()" does not work as I am doing paging
a least two solutions (please note that I do not check the identity classes members, so the following is the "spirit" of the solution (you miss the select clauses) ):
var result = _context.UserRoles.
Where(y => selectedRoles.Contains(y.RoleId)).
Select(y => y.User);
or
var result = _context.UserRoles.
Where(y => selectedRoles.Contains(y.RoleId)).
Select(y => y.UserId);
query = _context.Users.
Where(x => result.Contains(x.Id));
That said, assuming that there is no UserRoles table exposed in Identity (v2), you probably want:
userManager.Users.
Where(u => u.Roles.Any(r => selectecRoles.Contains(r.RoleId)));
Up to you to instanciate the userManager.
So for various reasons we need to send a large list of Ids to a EF6 query.
queryable.Where(x => list.Contains(x.Id));
is not ideal since it will create a huge were list.
So I was thinking, would it be possible some homehow to pass a table value parameter with the ids and get a IQueryable back that I can join against?
something like (Pseudo code)
var queryable = TableValueToIQueryable<MyTableValueType>(ids);
context.Set<MyEntity>().Join(queryable, x => x.Id, x.Value, (entity, id) => entity);
Is this possible somehow?
update: I have been able to use EntityFramework.CodeFirstStoreFunctions to execute a sql function and map the data to IQueryable<MyEntity>. it uses CreateQuery and ObjectParameters, can I use table value params somehow with ObjectParamters?
update2: Set().SqlQuery(...) will work with Table value parameters, but the resulting DbSqlQuery is not Joinable in SQL with a IQueryably so the result will be two connections and the join is done in memory
var idResult = Set<IdFilter>().SqlQuery("select * from GetIdFilter(#ids)", parameter);
var companies = idResult.Join(Set<tblCompany>(), x => x.Id, y => y.CompanyID, (filter, company) => company).ToList();
update3: ExecuteStoreQuery
((IObjectContextAdapter)ctx).ObjectContext.ExecuteStoreQuery<InvoicePoolingContext.IdFilter>("select * from dbo.GetIdFilter(#ids)", parameter)
.Join(ctx.Set<tblCompany>(), x => x.Id, y => y.CompanyID, (filter, company) => company).ToList();
Gives error:
There is already an open DataReader associated with this Command which
must be closed first.
Currently, to do a search using EF CodeFirst and a repository pattern, based on user input to multiple text boxes on an mvc search view/page, I do something like the following:
public PagedList<Entity1> PlayerUserSearch(Entity1SearchParameters searchParameters, int? pageSize, int? startEntity, Func<Entity1, object> sortOrder, bool sortDesc)
{
IQueryable<Entity1> query = from entities in this.DataContext.Entity1s.Include("Entity2List")
where entities.Entity2List.Any()
select entities;
if (searchParameters.Entity2PrimaryKeyId.HasValue)
query = query.Where(e => e.Id == searchParameters.Entity2PrimaryKeyId.Value);
if (searchParameters.HasStats.HasValue)
{
if (searchParameters.HasStats.Value)
query = query.Where(u => u.Entity2List.Any(e => e.Stat != null));
else
query = query.Where(u => u.Entity2List.Any(e => e.Stat == null));
}
if (searchParameters.Entity2OtherField.HasValue)
query = query.Where(u => u.Entity2List.Any(e => e.Event.Entity2OtherField == searchParameters.Entity2OtherField));
if (searchParameters.Entity2OtherField2.HasValue)
query = query.Where(u => u.Entity2List.Any(e => e.Event.Entity2OtherField2 == searchParameters.Entity2OtherField2));
if (searchParameters.Active.HasValue)
query = query.Where(e => e.Active == searchParameters.Active.Value);
return this.GetPageByStartEntity(pageSize.Value, startEntity.Value, query, sortOrder, sortDesc);
}
The problem with this is that for every time I add on another where that checks the child of Entity1 (Entity2) for a certain field, it takes on a new " AND EXISTS" clause to the sql statement generated, so that it is doing an exists and checking table Entity2 all over again for every different field checked, rather than doing a single EXISTS on Entity in the query, and checking all fields I tacked on to the query (i.e. EntityOtherField1 and EntityOtherField2). I haven't been able to find a better way to do a search based on user inputs than constantly checking for the input being there (add to the search parameters)) and then tacking on a new where to the current query. Can anyone tell me if there is a better way to do this? Thanks!
I think what you're looking for is using Specification Pattern.
var spec = new specification<Entity2>(s => true);
if (searchParameters.HasStats.Value)
{
spec = spec.And(e => e.Stat != null);
}
if (searchParameters.Entity2OtherField2.HasValue)
{
spec = spec.And(e => e.Event.Entity2OtherField2 == searchParameters.Entity2OtherField2);
}
query = query.Where(u => u.Entity2List.Any(spec));
Or I believe you can make it more standard by separating the filter logic and using spec.IsStatisfiedBy method.
A good framework for repository/specification implemetation on Entity Framework can be found here.
I am working on a new project and we are using Entity Framework and the dev lead would like to use lambda queries whenever possible. One thing we are having a hard time figuring out is how to select two columns specifically. Also how to select distinct. We have a table that has multiple entries for a vendor but we want to just get a list of vendors and load to a dictionary object. It fails because as written it is trying to add a key value that has already been added. Take the following query.
Dictionary<int, string> dict = new Dictionary<int, string>();
dict = GetWamVendorInfo().AsEnumerable()
.Where(x => x.vendor_name != null && x.vendor_id != null)
//.Select(x => x.vendor_id).Distinct()
.Take(2)
.ToDictionary(o => int.Parse(o.vendor_id.ToString()), o => o.vendor_name);
What I would like to do is select just vendor_id and vendor_name so we can get just the distinct records.
Any help would be greatly appreciated.
Thanks,
Rhonda
Use an anonymous type:
// earlier bit of query
.Select(x => new { VendorId = x.vendor_id, VendorName = x.vendor_name } )
.Distinct()
.ToDictionary(o => o.VendorId, o => o.VendorName);
I've removed the call to Take(2) as it wasn't clear why you'd want it - and also removed the parsing of VendorId, which I would have expected to already be an integer type.
Note that you should almost certainly remove the AsEnumerable call from your query - currently you'll be fetching all the vendors and filtering with LINQ to Objects. There's also no point creating an empty dictionary and then ignoring it entirely. I suspect your complete query should be:
var vendors = GetWamVendorInfo()
.Select(x => new { VendorId = x.vendor_id,
VendorName = x.vendor_name } )
.Distinct()
.ToDictionary(o => o.VendorId,
o => o.VendorName);
As an aside, you should ask your dev lead why he wants to use lambda expressions (presumably as opposed to query expressions) everywhere. Different situations end up with more readable code using different syntax options - it's worth being flexible on this front.
Just use an anonymous object:
var vendors = GetWamVendorInfo().AsEnumerable()
.Where(x => x.vendor_name != null && x.vendor_id != null)
.Select(new {x.vendor_id, x.vendor_name})
.Take(2)
That's it. You can now work with vendors[0].vendor_id, vendors[0].vendor_name, and so on.
I have an ASP.NET MVC application that displays data in a table format. I want to give my users the ability to search the table, so I take a text string and pass it into my service layer to construct a query using Linq to Entities.
I want to search a number of columns using the string. Some of the columns are integers (order ids), but the user doesn't care about integers and strings. They want to type '1200' and get any order with '1200' in the order number, or '1200' in the address.
The problem is that I can't find a way to construct a Linq-to-Entities query that results in SQL that looks like this:
select orderid, address from orders where orderid like '%1200%' or address like '%1200%'
Database context:
public DbSet<Person> Persons { get; set; }
public DbSet<Worker> Workers { get; set; }
public DbSet<WorkerSignin> WorkerSignins { get; set; }
The Persons and Workers tables are in a 1 to 0..1 relationship. If a worker record exists, a person record must also exist. They share the same ID. A worker record doesn't have to exist, however.
The Workers and WorkerSignins tables are related, but it's not enforced because of a client requirement. The Worker has an id-card with a barcode number on it (dwccardnum), but there may be discrepancies between cards issued and records in the DB, so I record all cards scanned in WorkerSignins, regardless of whether there is a matching record in the Workers table.
Here is the code I am working with:
allWSI = signinRepo.GetAllQ()
.Where(jj => jj.dateforsignin == date)
.Select(a => a);
if (!string.IsNullOrEmpty(search))
{
allWSI = allWSI
.Join(workerRepo.GetAllQ(), s => s.dwccardnum, w => w.dwccardnum, (s, w) => new { s, w })
.DefaultIfEmpty()
.Join(personRepo.GetAllQ(), oj => oj.w.ID, p => p.ID, (oj, p) => new { oj, p }).DefaultIfEmpty()
.DefaultIfEmpty()
.Where(jj => Convert.ToString(jj.oj.w.dwccardnum).Contains(search) ||
jj.p.firstname1.Contains(search) ||
jj.p.firstname2.Contains(search) ||
jj.p.lastname1.Contains(search) ||
jj.p.lastname2.Contains(search))
.Select(a => a.oj.s);
}
The GetAllQ() methods return an IQueryable() object.
The problem is on this line:
.Where(jj => Convert.ToString(jj.oj.w.dwccardnum).Contains(search) ||
I get this error:
LINQ to Entities does not recognize the method 'System.String ToString(Int32)' method, and this method cannot be translated into a store expression."
If I take out the convert, and try this:
.Where(jj => jj.oj.w.dwccardnum.Contains(search) ||
I get this error:
'int' does not contain a definition for 'Contains' and the best extension method overload 'System.Linq.ParallelEnumerable.Contains(System.Linq.ParallelQuery, TSource)' has some invalid arguments
So the question is...
How do I construct a Where clause to generate a like '%string%' and execute it against a integer column using Linq to Entities? (e.g. without using LINQ to SQL)
One option is to replace ...
jj => Convert.ToString(jj.oj.w.dwccardnum).Contains(search)
... by:
jj => SqlFunctions.StringConvert((decimal)jj.oj.w.dwccardnum).Contains(search)
SqlFunctions is a static class in namespace System.Data.Objects.SqlClient and I believe it only works with SQL Server. The weird cast to decimal is necessary because StringConvert doesn't have an overload for an int and without the cast the compiler complains that it cannot select the right overload unambiguously. (It has one for decimal? and one for double?.) But I just tested that the code above works indeed (with SQL Server and assuming dwccardnum is an int).
Try this
if (!string.IsNullOrEmpty(search))
{
int cardnum;
bool searchIsInt = int.TryParse(search, out cardnum);
allWSI = allWSI
.Join(workerRepo.GetAllQ(), s => s.dwccardnum, w => w.dwccardnum, (s, w) => new { s, w })
.DefaultIfEmpty()
.Join(personRepo.GetAllQ(), oj => oj.w.ID, p => p.ID, (oj, p) => new { oj, p }).DefaultIfEmpty()
.DefaultIfEmpty()
.Where(jj => (searchIsInt ? jj.oj.w.dwccardnum == cardnum : true) ||
jj.p.firstname1.Contains(search) ||
jj.p.firstname2.Contains(search) ||
jj.p.lastname1.Contains(search) ||
jj.p.lastname2.Contains(search))
.Select(a => a.oj.s);
}
Basically, you're first checking to see if the search is an int and then use it in your linq if it is.