I have following sql statement:
Select
tsl.Transaction_Id,
tsl.State_Id,
MAX(tsl."Timestamp")
from TransactionStatesLog tsl
group by tsl.Transaction_Id
How this statement can be translated to LINQ? I just want to select the whole row, where Timestamp is maximum of the group.
With this code i am able just to select TransactionId and max Timestamp from the group.
var states = (from logs in _context.TransactionStatesLog
group logs by new { logs.TransactionId } into g
select new
{
TransactionId = g.Key,
Timestamp = g.Max(x => x.Timestamp)
}).ToList();
I am working with ef core 3.1
Assuming you forgot to add tsl.State_Id to your SQL as grouping key as follows(otherwise that SQL does not work either):
Select
tsl.Transaction_Id,
tsl.State_Id,
MAX(tsl."Timestamp")
from TransactionStatesLog tsl
group by tsl.Transaction_Id, tsl.State_Id
If I understood you correctly you need to add StateId to grouping statement as well so that you will be able to select StateId and TransactionId.
So this should work:
var states = (from logs in _context.TransactionStatesLog
group logs by new { logs.TransactionId, logs.StateId } into g
select new
{
TransactionId = g.Key.TransactionId,
StateId = g.Key.StateId,
Timestamp = g.Max(x => x.Timestamp)
}).ToList();
See: Group by with multiple columns using lambda
Related
I am trying to generate a query that gets the number of entries with dates that occur within each given month. The SQL form of the query I wish to generate goes like this:
SELECT
DatePart(Year, [t1].[StartTime]) as Year,
DatePart(Month, [t1].[StartTime]) as Month,
Count(*) as 'Visits',
Sum([t1].[PageCount]) as 'Total Pageviews'
FROM
[MyDatabase] [t1]
GROUP BY
DatePart(Year, [t1].[StartTime]),
DatePart(Month, [t1].[StartTime])
I tried the following Linq2DB code:
var query = from table in dataContext.MyDataContext(tablePath)
group table by new { table.StartTime.Year, table.StartTime.Month} into grp
select new { Month = grp.Key.Month, Year = grp.Key.Year,
TotalVisitors = grp.Count(),
TotalPageviews = grp.Sum(table2 => table2.PageCount) };
but the SQL query being generated by this is
-- SqlServer.2008 --
SELECT
[t1].[StartTime],
Count(*) as [c1],
Sum([t1].[PageCount]) as [c2]
FROM
[MyDatabase] [t1]
GROUP BY
DatePart(Year, [t1].[StartTime]),
DatePart(Month, [t1].[StartTime]),
[t1].[StartTime]
Why is [t1].[StartTime] rather than the month or year? And why is it grouping by that extra [t1].[StartTime] at the end? How do I generate the SQL query I have above using Linq2DB?
Please use the following workaround with Sql.AsSql function. We are trying to fix this longstanding issue.
var query = from table in dataContext.MyDataContext(tablePath)
group table by new { Sql.AsSql(table.StartTime.Year), Sql.AsSql(table.StartTime.Month)} into grp
select new { Month = Sql.AsSql(grp.Key.Month), Year = Sql.AsSql(grp.Key.Year),
TotalVisitors = grp.Count(),
TotalPageviews = grp.Sum(table2 => table2.PageCount) };
I have a sql like below:
select catalog,queryname,count(*) from table group by catalog,queryname
Now how can I write it in linq and get the count?
thanks
var g =
from record in table
group record by new
{
record.catalog, record.queryname
} into mygroup
select new
{
catalog = key.catalog,
queryname = key.queryname
count = myGroup.Count()
};
I want to return all the values from my "Categories" table and join that to my "CategorySelections" table to display all the categories and whether the specified user selected them or not.
A complication (which I don't know how to deal with in LINQ) is that the user could have changed his selection/deselection of a particular category over time...each change would have logged in the "CategorySelections" table with a date stamp.
I am after the last selection status.
The following SQL query does what I want:
SELECT cs.UserId, c.CategoryId, m.MaxDate, cs.IsSelected
FROM [myDB].[dbo].[Categories] c
LEFT JOIN [myDB].[dbo].[CategorySelections] cs
ON c.CategoryID = cs.CategoryID AND cs.UserID = 7
INNER JOIN
(
SELECT UserId, CategoryId, Max(CreatedOn) as MaxDate
FROM [myDB].[dbo].[CategorySelections]
GROUP BY UserId, CategoryId
) m
ON cs.UserID = m.UserID AND cs.CategoryID = m.CategoryID AND cs.CreatedOn = m.MaxDate
ORDER BY cs.CategoryI
I need some help getting this done in LINQ.
Below is my attempt, which returns all the selections instead of just the last per category.
var query = from c in db.Category
join cs in db.CategorySelection.Where(x => x.UserID == WebSecurity.CurrentUserId)
on c.CategoryID equals cs.CategoryID into JoinedCategory
from cs in JoinedCategory.DefaultIfEmpty()
select new Selection() { CategoryID = c.CategoryID, CategoryName = c.CategoryName ,IsSelected = cs != null ? cs.IsSelected : false }
I am working in MVC; the "new Selection()" refers to my Model
You may add WHERE statement:
where cs.CreatedOn == CategorySelections.Where(t => t.CategoryId == cs.CategoryId).Max(r => r.CreatedOn)
How to translate query like "select 1, 2" (i.e. without FROM clause) to LINQ statement?
Thanks!
I need to get permissions for a set of user groups. In SQL it looks like
SELECT *
FROM Permission p
INNER JOIN (SELECT GroupID
FROM [Group]
UNION ALL
SELECT 555) AS g
ON (g.GroupID = p.GroupID)
In my case I need to programmatically add a certain code instead "555". I wouldn't like to write special SQL function for that.
I guess you just want to create an anonymous type
var anonymous = new { Column1 = 1, Column2 = 2 };
Edit - Based on Comments
Depending on what your Select projection is you could do something simple like this:
If it is a Int:
var query = (from per in context.permissions
select per).AsEnumerable()
.Concat( new int[] { 1, 2 });
If it is a 'Class'
var query = (from per in context.permissions
select per).AsEnumerable()
.Concat(new CustomClass[]
{
new CustomClass()
{
Prop1= 1
},
}
);
You could also change .Concat to .Union
Why do you need this to be linq?
var numbers = new int[] { 1, 2 };
I suppose
var numbers = Enumerable.Range(1,2);
How can I write a linq to entities query that includes a having clause?
For example:
SELECT State.Name, Count(*) FROM State
INNER JOIN StateOwner ON State.StateID = StateOwner.StateID
GROUP BY State.StateID
HAVING Count(*) > 1
Any reason not to just use a where clause on the result?
var query = from state in states
join stateowner in stateowners
on state.stateid equals stateowner.stateid
group state.Name by state.stateid into grouped
where grouped.Count() > 1
select new { Name = grouped.Key, grouped.Count() };
I believe you can use a GroupBy followed by a Where clause and it will translate it as a Having. Not entirely sure though.
If you want to compare a variable that is not in the group by (Ex: age), then it would be:
var duplicated = (
from q1 in db.table1
where (q1.age >= 10 )
group q1 by new { q1.firstName, q1.lastName } into grp
where (grp.Count() > 1 )
select new
{
firstName= grp.Key.firstName,
lastName = grp.Key.lastName,
}
);