We Keep Coding

Linq query equivalent of exists with multiple conditions

I've been using linq for a little while now but haven't come across this situation anywhere and my google-fu let me down.
Basically I have two data sets which I did not define and now have to use to return data.
class Header
{
string COMPANY_CODE
string REFERENCE_NBR
string REFERENCE_DUPLICATE
...
}
class Line
{
string COMPANY_CODE
string REFERENCE_NBR
string REFERENCE_DUPLICATE
string STOCK_CODE
string DESCRIPTION
...
}
From a database perspective they join like this
select *
from Header
inner join Line
on header.COMPANY_CODE = Line.COMPANY_CODE
and header.REFERENCE_NBR = Line.REFERNCE_NBR
and header.REFERENCE_DUPLICATE = LINE.REFERENCE_DUPLICATE
and have a 1:Many relationship.
I'm implementing a search feature for a listing that is meant to find any Lines that with a value in STOCK_CODE or DESCRIPTION that matches a given search term. I have seen a couple of methods of joining using a linq query but because of the multiple join conditions I'm a bit lost and have not found any examples of what I'm trying to do.
If I were to write the statement I am trying to get in lamda/linq in SQL it would be:
declare #searchtxt nvarchar(max) = 'test'
Select *
from header h
where exists (
select *
from Line l
where
(
l.stock_code like '%'+#searchtxt+'%'
or l.description like '%'+#searchtxt+'%'
)
and h.COMPANY_CODE = l.COMPANY_CODE
and h.REFERENCE_NBR = l.REFERENCE_NBR
and h.REFERENCE_DUPLICATE = l.REFERENCE_DUPLICATE
)
Any help would be appreciated!

Perhaps, this?
var result = header.Where(h =>
Line.Any(l => (l.stock_code.Contains(searchtxt)
|| l.description.Contains(searchtxt))
&& h.COMPANY_CODE == l.COMPANY_CODE
&& h.REFERENCE_NBR == l.REFERENCE_NBR
&& h.REFERENCE_DUPLICATE == l.REFERENCE_DUPLICATE));

This is a traditional LINQ query for better understanding,
string searchtext = "";
var result = (from h in context.Headers
join l in context.Lines on new { h.COMPANY_CODE, h.REFERENCE_DUPLICATE, h.REFERENCE_NBR } equals new { l.COMPANY_CODE, l.REFERENCE_DUPLICATE, l.REFERENCE_NBR }
where l.STOCK_CODE.Contains(searchtext) || l.DESCRIPTION.Contains(searchtext)
select new
{
COMPANY_CODE = h.COMPANY_CODE,
STOCK_CODE = l.STOCK_CODE
//You can select more fields from "h" and "l"
}).ToList();
Edit:
string searchtext = "";
var result = (from h in context.Headers
join l in context.Lines on new { h.COMPANY_CODE, h.REFERENCE_DUPLICATE, h.REFERENCE_NBR } equals new { l.COMPANY_CODE, l.REFERENCE_DUPLICATE, l.REFERENCE_NBR }
where l.STOCK_CODE.Contains(searchtext) || l.DESCRIPTION.Contains(searchtext)
select h
)
.GroupBy(x => new { x.COMPANY_CODE, x.REFERENCE_DUPLICATE, x.REFERENCE_NBR })
.Select(x => x.First())
.ToList();

db.Header.Join(
db.Line,
h => new { h.COMPANY_CODE, h.REFERENCE_NBR, h.REFERENCE_DUPLICATE },
l => new { l.COMPANY_CODE, l.REFERENCE_NBR, l.REFERENCE_DUPLICATE },
(h, l) => new
{
Header_COMPANY_CODE = h.COMPANY_CODE,
Header_REFERENCE_NBR = h.REFERENCE_NBR,
Header_REFERENCE_DUPLICATE = h.REFERENCE_DUPLICATE,
Line_Company_Code = l.COMPANY_CODE,
Line_REFERENCE_NBR = l.REFERENCE_NBR,
Line_REFERENCE_DUPLICATE = l.REFERENCE_DUPLICATE,
Line_STOCK_CODE = l.STOCK_CODE,
Line_DESCRIPTION = l.DESCRIPTION
}
)
.Where(w => w.Line_STOCK_CODE.Contains(searchText) || w.Line_DESCRIPTION.Contains(searchText))
.ToList();

Linq to SQL grouping with embedded lists causes too many queries

I am developing a query to grab and join some SQL tables in C# and am having some trouble with grouping and enumerables within the dataset. My query is below. This gives me the data in the format I'm looking for, but it takes way too long when I try to add the enumerated list as indicated below. When I look under the hood I can see it is executing way too many SQL queries. I'd like to get it to just one. Using LinqPad:
void Main()
{
var nightlyRuns = (from a in LoadTestSummaries
join b in LoadTestTestSummaryData
on a.LoadTestRunId equals b.LoadTestRunId
where a.TargetStack == "LoadEnv" &&
a.TestGuid != null &&
a.StartTime != null &&
a.LoadTestRunId != null
orderby a.StartTime
group new {a, b} by new
{
a.TestGuid,
a.Name,
a.Description,
a.StartTime,
a.Duration,
a.NumAgents,
a.NumHosts,
a.PassFail,
a.ResultsFilePath,
a.Splunk
}
into g
let scenarioStart = g.Min(s => s.a.StartTime) ?? g.Min(s => s.a.DateCreated)
let testCases = g.Select(s => s.b)
orderby scenarioStart
select new
{
TestGuid = g.Key.TestGuid,
ScenarioRun = new
{
Name = g.Key.Name,
Description = g.Key.Description,
StartTime = scenarioStart,
Duration = g.Key.Duration,
NumAgents = g.Key.NumAgents,
NumHosts = g.Key.NumHosts,
Result = g.Key.PassFail,
ResultsFilePath = g.Key.ResultsFilePath,
SplunkLink = g.Key.Splunk,
// PROBLEM: Causes too many queries:
TestRuns = from t in testCases select t.TestCaseId
}
}).ToLookup(g => g.TestGuid, g => g.ScenarioRun);
nightlyRuns["ba593f66-695f-4fd1-99c3-71253a2e4981"].Dump();
}
The "TestRuns" line is causing the excessive queries. Any idea what I am doing wrong here?
Thanks for any insight.

Tough answer to test but I think we can avoid the grouping and multiple queries with something like this: (https://msdn.microsoft.com/en-us/library/bb311040.aspx)
var nightlyRuns = (from a in LoadTestSummaries
join b in LoadTestTestSummaryData
on a.LoadTestRunId equals b.LoadTestRunId
where a.TargetStack == "LoadEnv" &&
a.TestGuid != null &&
a.StartTime != null &&
a.LoadTestRunId != null
into testGroup
select new
{
TestGuid = a.TestGuid,
ScenarioRun = new
{
Name = a.TestGuid,
Description = a.Description,
StartTime = a.StartTime ?? a.DateCreated,
Duration = a.Duration,
NumAgents = g.Key.NumAgents,
NumHosts = a.NumHosts,
Result = a.PassFail,
ResultsFilePath = a.ResultsFilePath,
SplunkLink = a.Splunk,
// PROBLEM: Causes too many queries:
TestRuns =testGroup
}
}).OrderBy(x=>x.StartTime).ToLookup(x => x.TestGuid, x => x.ScenarioRun);
nightlyRuns["ba593f66-695f-4fd1-99c3-71253a2e4981"].Dump();

LINQ with multiple columns and operations

i'm trying to do this on LINQ:
select p.ProductID, p.ArticleCode, SUM((case h.OperationType when 0 then 1 else -1 end) * h.[Count])
from Products p
inner join StockItems s on p.ArticleCode = s.ArticleCode
inner join StockHistorical h on s.ArticleID = h.ArticleID
where h.[Date] < '23/08/2013 11:30:00'
group by p.ProductID, p.ArticleCode
I have this:
var products = (from p in e.Products
join s in e.StockItems on p.ArticleCode equals s.ArticleCode
join h in e.StockHistoricals on s.ArticleID equals h.ArticleID
where h.Date < DateTime.Parse("23/08/2013 11:30:00")
group p by p.ProductID into StockResult
select new { });
Anyone know how can i do the
SUM((case h.OperationType when 0 then 1 else -1 end) * h.[Count])
with LINQ?
Thanks!
I forgot to say that the problem is the "group by", because i can't access the OperationType property in the StockResult group.
Solved! The key is the:
let combined = new
{
Product = p,
HistoryCount = (h.OperationType == 0 ? 1 : -1) * h.Count
}

...
let combined = new
{
Product = p,
HistoryCount = (h.OperationType == 0 ? 1 : -1) * h.Count
}
group combined by combined.Product.ProductID into StockResult
select new
{
ProductID = StockResult.Key,
Total = StockResult.Sum(x => x.HistoryCount)
}

It's not a direct translation, but I'd go with:
(h.Count(his => his.OperationType != 0)
- h.Count(his => his.OperationType == 0))
* h.Count()
It should be functionally equivalent, and seems to better represent what you're trying to do.

You are trying to do something like this?
I am not sure this is going to work.
select new
{
S=Sum(((h.OperationType)!=0?-1:1)*h.Count)
}

linq to object query to exclude 0

I have a Linq query as below.
var DataSource = from m in product
select new { Class = m.Class, Id = (new Make(m.Id)).ProductName};
I instantiate a class called Make and fetch the ProductName based on the Id.
Some of the Id's are 0.
Is there a way I can exclude all the Id's that are 0?

Sure, just do this:
var DataSource = from m in product
where m.Id != 0
select new
{
Class = m.Class,
Id = (new Make(m.Id)).ProductName
};

You're looking for the where clause:
where m.Id != 0

var DataSource = from m in product
where m.Id != 0
select new
{
Class = m.Class,
Id = (new Make(m.Id)).ProductName
};

Try this:
var DataSource = from m in product
where m.Id != 0
select new { Class = m.Class, Id = (new Make(m.Id)).ProductName};

How to combine 2 linq statments with groupby clause into 1

I was wondering if i can consolidate below 2 linq statments into 1 statment. I am sure it should be possible, but various attempts i am unable to manage.
var prevProvisionsBySubBook = (from provision in prevProvisions
group provision by provision.SubBook
into subBookGrouping
select
new
{
Key = subBookGrouping.Key,
Value = subBookGrouping.Sum(t => t.ProvisionUSD)
});
var currentProvisionsBySubBook
= (from provision in currentProvisions
group provision by provision.SubBook
into subBookGrouping
select new
{
Key = subBookGrouping.Key,
Value = subBookGrouping.Sum(t => t.ProvisionUSD)
});
var adjustmentChangeBySubBook
= (from current in currentProvisionsBySubBook
select new
{
Key = current.Key,
Value = current.Value
- (prevProvisionsBySubBook.Any() ? prevProvisionsBySubBook.Where(t => t.Key == current.Key).Single().Value : 0)
});
any help would be apprecaited.

You can do it like this:
var adjustmentChangeBySubBook =
from provision in
(from currentProvision in currentProvisions select new
{
currentProvision.SubBook,
CurrentUSD = currentProvision.ProvisionUSD,
PreviousUSD = 0
}).Concat
(from prevProvision in prevProvisions select new
{
prevProvision.SubBook,
CurrentUSD = 0,
PreviousUSD = prevProvision.ProvisionUSD
})
group provision by provision.SubBook into subBookGrouping select new
{
subBookGrouping.Key,
Value = subBookGrouping.Sum(t => t.CurrentUSD - t.PreviousUSD)
};