SQL query to LINQ conversion with nested select statements - c#

I want to convert the following query to LINQ:
SELECT LV.* FROM LowerVehicles LV
INNER JOIN (Select VSerial,MAX(updatedOn) MaxUpdatedOn from LowerVehicles group by vserial) LVG
ON LV.VSerial = LVG.VSerial AND LV.updatedOn = LVG.MaxUpdatedOn


Not knowing your entities classes, here is an approximation. You can use query syntax or fluent syntax. Sometimes one is preferable over the other, and in the case of joins and grouping I prefer to use query syntax.
QUERY SYNTAX
var query = from LV in LowerVehicles
join LVG in (
from r in LowerVehicles
group r by r.vserial into g
select new {VSerial = g.Key, MaxUpdatedOn = g.Max(t => t.updatedOn)})
on LV.VSerial equals LVG.Vserial
and LV.updatedOn equals LVG.MaxUpdatedOn
select LV;
FLUENT SYNTAX
var lvg = LowerVehicles.GroupBy(t => t.vserial)
.Select(g => new {
VSerial = g.Key,
MaxUpdatedOn = g.Max(t => t.updatedOn)
});
var query = LowerVehicles.Join(
lvg,
a => new { a.VSerial, a.updatedOn },
b => new { b.VSerial, b.MaxUpdatedOn },
(a, b) => new { LV = a, LVG = b}
)
.Select(t=> t.LV);


Something like this?
Something.LowerVehicles
.Join(something.LowerVehicles.Select(y => new { y.VSerial, updatedOn = y.updatedOn.Max() }).GroupBy(z => z.VSerial),
x => new { x.VSerial, x.updatedOn },
lvg => new { lvg.VSerial, lvg.updatedOn },
(x, y) => x)

Related

Simplify many LINQ Join

I have a rather complex LINQ query but that is the point of the question
var result = await _context.TblBreakpoints
.GroupBy(b => new { b.BpgroupId, b.ResultType, b.DrugId, b.Susceptible, b.LowIntermediate, b.Intermediate, b.Resistant})
.Join(_context.TblBreakpointgroups,
bg => bg.Key.BpgroupId,
g => g.BpgroupId,
(bg, g) => new
{
GroupId = bg.Key.BpgroupId,
DrugId = bg.Key.DrugId,
Susceptible = bg.Key.Susceptible,
LowIntermediate = bg.Key.LowIntermediate,
Intermediate = bg.Key.Intermediate,
Method = bg.Key.ResultType,
Resistant = bg.Key.Resistant,
StandardId = g.BpstandardId,
GroupName = g.BpgroupName,
Count = bg.Count(),
})
.Join(_context.TblBreakpointStandards,
i => i.StandardId,
j => j.BpstandardId,
(i, j) => new
{
Standard = j.Bpstandard,
GroupId = i.GroupId,
GroupName = i.GroupName,
Count = i.Count,
Method = i.Method,
DrugId = i.DrugId,
Susceptible = i.Susceptible,
LowIntermediate = i.LowIntermediate,
Intermediate = i.Intermediate,
Resistant = i.Resistant
})
.Join(_context.TblDrugs,
i => i.DrugId,
j => j.DrugId,
(i, j) => new
{
DrugName = j.DrugName,
Standard = i.Standard,
GroupId = i.GroupId,
GroupName = i.GroupName,
Count = i.Count,
Method = i.Method,
Susceptible = i.Susceptible,
LowIntermediate = i.LowIntermediate,
Intermediate = i.Intermediate,
Resistant = i.Resistant
})
.Join(_context.TblBreakpointgroupmembers,
i => i.GroupId,
j => j.BpgroupId,
(i, j) => new
{
OrganismId = j.OrganismId,
Standard = i.Standard,
GroupId = i.GroupId,
GroupName = i.GroupName,
Count = i.Count,
Method = i.Method,
DrugName = i.DrugName,
Susceptible = i.Susceptible,
LowIntermediate = i.LowIntermediate,
Intermediate = i.Intermediate,
Resistant = i.Resistant
})
.Join(_context.TblOrganismNames,
i => i.OrganismId,
j => j.OrganismId,
(i, j) => new BreakpointSummary
{
OrganismName = j.OrganismName,
Standard = i.Standard,
GroupName = i.GroupName,
Count = i.Count,
Method = i.Method,
DrugName = i.DrugName,
Susceptible = i.Susceptible,
LowIntermediate = i.LowIntermediate,
Intermediate = i.Intermediate,
Resistant = i.Resistant
})
.ToListAsync().ConfigureAwait(false);
From the query with each Join I keep passing the previous values and add the value(s) that come from the join. It is already tedious with just 5 joins, it would get even more so with more joins. Is there a better way that I am missing?
I think the equivalent SQL is
WITH bpg (BPGroupId, ResultType, DrugId, Susceptible, LowIntermediate, Intermediate, Resistant, Count)
AS (
SELECT BPGroupId, ResultType, DrugId, Susceptible, LowIntermediate, Intermediate, Resistant, COUNT(BPGroupId)
FROM dbo.tbl_Breakpoint a
GROUP BY BPGroupId,
ResultType,
DrugId,
Susceptible,
LowIntermediate,
Intermediate,
Resistant
)
SELECT a.BpgroupName, b.BPStandard, c.DrugName, e.OrganismName, CTE.ResultType, CTE.Susceptible, CTE.LowIntermediate, CTE.Intermediate, CTE.Resistant, CTE.Count
FROM dbo.tbl_breakpointgroup a
INNER JOIN bpg CTE ON a.BPGroupId = CTE.BPGroupId
INNER JOIN tbl_BreakpointStandard b ON b.BPStandardId = a.BPStandardId
INNER JOIN tbl_Drug c ON c.DrugID = CTE.DrugId
INNER JOIN tbl_breakpointgroupmember d ON d.BPGroupId = CTE.BPGroupId
INNER JOIN tbl_OrganismName e ON e.OrganismId = d.OrganismId
WHERE a.BPGroupId = CTE.BPGroupId

In general when using manual joins in LINQ, it's better to use the query syntax since it provides range variables (which correspond to table/query aliases in SQL) transparency. e.g. (in pseudo code)
from a in queryA
join b in queryB on ... // can use any field from a
join c in queryC on ... // can use any field from a and b
join d in queryD on ... // can use any field from a, b and c
...
select new
{
// can use any field for a, b, c, d etc.
}
The same with method syntax is a bit more complicated, but the principle is to wrap the previous "variables" in simple tuple like anonymous types until you get to the final projection, e.g. (in pseudo code)
queryA
.Join(queryB, a => {a fields}, b => {b fields), (a, b) => new { a, b }) // first result
.Join(queryC, r => {r.a, r.b fields), c => { c fields }, (r, c) => new { r.a, r.b, c } // second result
.Join(queryD, r => {r.a, r.b, r.c fields), d => { d fields }, (r, d) => new { r.a, r.b, r.c, d } // third result
...
.Select(r => new { r.a, r.b, r.c, r.d... fields });
Applying it to your example, the corresponding query syntax could be like (note that sub queries inside can use whatever syntax is appropriate):
var query =
from cte in _context.TblBreakpoints
.GroupBy(b => new { b.BpgroupId, b.ResultType, b.DrugId, b.Susceptible, b.LowIntermediate, b.Intermediate, b.Resistant})
.Select(g => new
{
g.Key.BpgroupId, g.Key.ResultType, g.Key.DrugId, g.Key.Susceptible, h.Key.LowIntermediate, g.Key.Intermediate, g.Key.Resistant,
Count = g.Count(),
})
join a in _context.TblBreakpointgroups on cte.BpgroupId equals a.BpgroupId
join b in _context.TblBreakpointStandards on a.BpstandardId equals b.BpstandardId
join c in _context.TblDrugs on cte.DrugId equals c.DrugId
join d in _context.TblBreakpointgroupmembers on cte.BpgroupId equals d.BpgroupId
join e in _context.TblOrganismNames on d.OrganismId equals e.OrganismId
select new BreakpointSummary
{
OrganismName = e.OrganismName,
Standard = b.Bpstandard,
GroupName = a.BpgroupName,
Count = cte.Count,
Method = cte.ResultType,
DrugName = d.DrugName,
Susceptible = cte.Susceptible,
LowIntermediate = cte.LowIntermediate,
Intermediate = cte.Intermediate,
Resistant = cte.Resistant,
};
You can convert it to method syntax using the aforementioned rules, but for me it doesn't worth the effort.

Lambda syntax. How join by 1 to many keys

i'm trying rewrite query from native sql to linq lambda syntax (not linq query syntax)
LINQ (not work)
var result = _uow.Repository<TableA>().Get().AsNoTracking()
.GroupJoin(
_uow.Repository<TableB>().Get().AsNoTracking(),
a => new { a.TabNotesCodeId, a.TabLabelCodeId },
b => b.ElementNameId
(b, a) => new SubSection
{
SubSectionName = b.CustomValue ?? a.TabLabelCodeId,
SubSectionNote = b.CustomValue ?? a.TabLabelCodeId,
})
.Where(a => a.ResourceId == 1);
SQL
SELECT [SubSectionName] = ISNULL(B.CUSTOMVALUE,A.TABLABELCODEID),
[SubSectionNote] = ISNULL(B.CUSTOMVALUE,A.TABNOTESCODEID)
FROM TableA as A LEFT JOIN
(SELECT CUSTOMVALUE, ELEMENTNAMEID FROM TableB WHERE DISPLAYSETTINGID = 1) as B
ON B.ELEMENTNAMEID IN ( A.TABNOTESCODEID, A.TABLABELCODEID)
WHERE A.RESOURCEID = 1
Q How to rewrite sql ON B.ELEMENTNAMEID IN ( A.TABNOTESCODEID, A.TABLABELCODEID) to lambda syntax
...
a => new { a.TabNotesCodeId, a.TabLabelCodeId },
b => b.ElementNameId
....
(doesn't work)

Normally I would suggest following my SQL conversion rules, but this is sufficiently complex I don't think it would help.
To use query comprehension syntax on a non-equijoin that is a left join, it seems easiest to use lambda syntax to express the join conditions, so I just combined the sub-query with the join conditions:
var ans = from A in TableA
where A.ResourceID == 1
from B in TableB.Where(b => b.DisplaySettingID == 1).Where(b => b.ElementNameID == A.TabNotesCodeID || b.ElementNameID == A.TabLabelCodeID).DefaultIfEmpty()
select new {
SubSectionName = (B.CustomValue ?? A.TabLabelCodeID),
SubSectionNote = (B.CustomValue ?? A.TabNotesCodeID)
};
The lambda equivalent of multiple from clauses to generate a cross join is SelectMany, so converting into lambda syntax:
var ans2 = TableA.Where(a => a.ResourceID == 1)
.SelectMany(a => TableB.Where(b => b.DisplaySettingID == 1).Where(b => b.ElementNameID == a.TabNotesCodeID || b.ElementNameID == a.TabLabelCodeID)
.DefaultIfEmpty(),
(a, b) => new {
SubSectionName = (b.CustomValue ?? a.TabLabelCodeID),
SubSectionNote = (b.CustomValue ?? a.TabNotesCodeID)
}
);

After countless experiments i've found out solution:
_uow.Repository<TableA>().Get().AsNoTracking()
.GroupJoin(
_uow.Repository<TableB>().Get().AsNoTracking().Where(b => b.DisplaySettingId == 1),
a => new { note = a.TabNotesCodeId, label = a.TabLabelCodeId },
b => new { note = b.ElementNameId, label = b.ElementNameId },
(a, b) => new { a,b })
.Where(joinTables => joinTables.a.ResourceId == 1)
.SelectMany(
joinTables => joinTables.b.DefaultIfEmpty(),
(joinTables, b) => new SubSection()
{
LayoutTab = joinTables.a.LayoutTab,
SubSectionName = b.CustomValue ?? joinTables.a.TabLabelCodeId,
SubSectionNote = b.CustomValue ?? joinTables.a.TabNotesCodeId
});

Join and Group By using Lambda Expression

Query
var grpby4 = from u in dtEmp.AsEnumerable()
join v in dtDept.AsEnumerable() on u.Field<int>("DepartmentID") equals v.Field<int>("DepartmentID")
group u by v.Field<string>("DeptName") into g
select new { DeptName = g.Key, Records = g };
How write the same Query using Lambda Expression?

Using this handy webpage I get
dtEmp.AsEnumerable()
.Join(dtDept.AsEnumerable(),
u => u.Field<int>("DepartmentID"),
v => v.Field<int>("DepartmentID"),
(u, v) => new { u, v })
.GroupBy(τ0 => τ0.v.Field<string>("DeptName"), τ0 => τ0.u)
.Select(g => new { DeptName = g.Key, Records = g })

Translate SQL to lambda LINQ with GroupBy and Average

I spend a few hours trying to translate simple SQL to lambda LINQ
SELECT ID, AVG(Score) FROM myTable
GROUP BY ID
Any idea?

from t in myTable
group t by new {
t.ID
} into g
select new {
Average = g.Average(p => p.Score),
g.Key.ID
}
or Lambda
myTable.GroupBy(t => new {ID = t.ID})
.Select (g => new {
Average = g.Average (p => p.Score),
ID = g.Key.ID
})

The equivalent in Linq-to-Objects would be something like the below.
var results = from row in myTable
group row by row.Id into rows
select new
{
Id = rows.Key,
AverageScore = rows.Average(row => row.Score)
};
It's only slightly different for an ORM like entity framework. Namely, you would need to go through the data context or an appropriate DbSet/ObjectSet.

var _result = from a in myTable
group a by a.ID into g
select new
{
ID = g.Key.ID,
AverageResult = g.Average(x => x.Score)
}

Left outer join of datatable with linq expression?

I have two DataTables:
DataTable dtFields = new DataTable("tmpFieldTable");
dtFields.Columns.Add("FieldID");
dtFields.Columns.Add("CDGroupID");
dtFields.Columns.Add("CDCaption");
dtFields.Columns.Add("fldIndex");
and
DataTable dtCDGroup = new DataTable("tmpCDGroup");
dtCDGroup.Columns.Add("CDGroupID");
dtCDGroup.Columns.Add("Name");
dtCDGroup.Columns.Add("Priority");
I am using following LINQ expression to join these tables:
var resultList = dtFields.AsEnumerable()
.Join(dtCDGroup.AsEnumerable(),
fieldList => fieldList.Field<string>("CDGroupID"),
cd => cd.Field<string>("CDGroupID"),
(fieldList, cd) => new
{
FieldID = fieldList.Field<string>("FieldID"),
CdGroup = cd.Field<string>("Name"),
CDCaption = fieldList.Field<string>("CDCaption"),
Priority = (cd.Field<string>("Priority") ?? "99"),
fldIndex = fieldList.Field<string>("fldIndex").ToString()
})
.OrderBy(result => result.Priority)
.ThenBy(result => result.fldIndex);
How can I perform left outer join with these DataTables?

Left join are not natively offered using lambda expression. but you can achieve this using GroupJoin and SelectMany method.
var result = new string[] { "A", "B" }.GroupJoin(new string[] { "B" }, l => l, r => r, (l, r) => new { l, r }).ToList();
Console.WriteLine(result.Count());
this will result result[0].r.Count() = 0 and result[1].r.Count() = 1.
Hope this will help.
Regards.

Just add DefaultIfEmpty() to your table:
resultList = dtFields.AsEnumerable()
.Join(dtCDGroup.AsEnumerable().DefaultIfEmpty(),
fieldList => fieldList.Field<string>("CDGroupID"),
cd => cd.Field<string>("CDGroupID"),
(fieldList, cd) => new
{
FieldID = fieldList.Field<string>("FieldID"),
CdGroup = cd.Field<string>("Name"),
CDCaption = fieldList.Field<string>("CDCaption"),
Priority = (cd.Field<string>("Priority") ?? "99"),
fldIndex = fieldList.Field<string>("fldIndex").ToString()
})
.OrderBy(result => result.Priority)
.ThenBy(result => result.fldIndex);

Categories