How to create Linq2Sql query that will group records from linked table and calculate 2 count fields
-
01-06-2021 - |
Question
Here I found how to join tables using Linq2Sql and count amount of linked records LINQ - Left Join, Group By, and Count
I've implemented it and it works ok for me: the following expression
var v = from c in db.GetTable<Country>()
join t0 in db.GetTable<Team>() on c.Id equals t0.CountryId into t1
from team in t1.DefaultIfEmpty()
group team by c.Id into teamsGrouped
select new CountryTeamsInfo
{
CountryId = teamsGrouped.Key,
TeamsTotal = teamsGrouped.Count(),
// TeamsWithoutOwnerFree = teamsGrouped.Count(t => t.OwnerId==0)
}
;
List<CountryTeamsInfo> res = v.ToList();
generates the following query:
SELECT c.Id, Count(*) as c1
FROM countries c
LEFT JOIN teams t1 ON c.Id = t1.Country
GROUP BY c.Id
In fact I need also to count those linker records that has OwnerId field non-equal to 0.
It looks like I should just uncomment that line in linq expression (TeamsWithoutOwnerFree = teamsGrouped.Count(t => t.OwnerId==0)), but that doesn't work, attempt to execute that causes an error:
The given key was not present in the dictionary
The query don't come to SQL log file, and I can't examine it in the debugger.
What should be a proper way to count those lines from 'teams' table that meet additional criteria.
P.S. if it matters, I use C# 4.0, MySql 5.1 and BLToolkit 4.1
Solution 3
Here is query that works in my environment:
from c in db.GetTable<Country>()
where c.Allowed
select new CountryTeamsInfo
{
CountryId = c.Id,
TeamsTotal = db.GetTable<Team>().Count(t => t.CountryId == c.Id && t.Allowed),
TeamsHasOwner = db.GetTable<Team>().Count(t => t.CountryId == c.Id && t.Allowed && t.OwnerId != 0),
}
Not the best solution (I need to repeat teams selection criteria t.Allowed in each sub-query), but still works well and generated SQL is good.
OTHER TIPS
Maybe try using GroupJoin()
for your query:
var result = db.GetTable<Country>()
.GroupJoin(db.GetTable<Team>(),
c => c.Id,
t => t.CountryId,
(country, teams) => new
{
CountryId = country.Id,
TeamsTotal = teams.Count(),
TeamsWithoutOwnerFree = teams.Count(t => t.OwnerId == 0)
})
.ToList();
Due to RePierre help I was able to find the proper query:
var v = db.GetTable<Country>().Where(country => country.Allowed)
.GroupJoin(
db.GetTable<Team>(),
country => country.Id,
team => team.CountryId,
(country, teams) => new CountryTeamsInfo
{
CountryId = country.Id,
TeamsTotal = teams.Count(),
TeamsWithoutOwnerFree = teams.Count(t => t.OwnerId != 0),
}
).GroupJoin(
db.GetTable<Team>().Where(te=>te.OwnerId==0),
cti => cti.CountryId,
team => team.CountryId,
(cti, teams) => new CountryTeamsInfo
{
CountryId = cti.CountryId,
TeamsTotal = cti.TeamsTotal,
TeamsWithoutOwnerFree = teams.Count(t => t.OwnerId != 0),
}
)
;
My concern is it uses 2 sub-queries... will try to optimize that. Any ideas would be great
P.S. Actually, generated SQL query looks ugly as well:
SELECT
cti.Id as Id1,
cti.c1 as c11,
(
SELECT
Count(*)
FROM
teams te
WHERE
cti.Id = te.Country AND te.User = 0
) as c2
FROM
(
SELECT
country.Id,
(
SELECT
Count(*)
FROM
teams t1
WHERE
country.Id = t1.Country
) as c1
FROM
countries country
WHERE
country.allow