Develop Reference

Linq help - sub query gives null pointer exception

I writing a email system where we have a table of users "tblUsers" and a table of messages. A user can have many messages (from other users in tblusers) in his or her inbox (one:many).
In tblUsers table, I have a column called ImageURL (string) that contains the URL to the user's avatar. In this case, I'm looping through the messages in an inbox belonging to a user and what I'm trying to do is, once I get the message, walk up the tree to the tblUser and get the value in the ImageURL column for the owner of that message as marked "SenderAvatar" below.
Here's what I tried. The problem is that the sub linq for SenderAvatar below is throwing a nullpointer exception even though I have confirmed that there is a value for ImageURL (this is dev so there's only three users). Somehow my logic and linq's logic is at odds here. Can someone please help? Thanks!
Edit
I found two bugs. The first bug is Dzienny pointed me to the right direction where I was comparing apples and oranges. The second bug is FromUserId = ux.tblUserId, where I'm setting the current user id to FromUserId Guys, thank you for all your help on this.
public List<UserInboxMsg> GetUserInboxMsg(IKASLWSEntities conx, int userid)
{
var u = (from m in conx.tblUsers where m.Id == userid select m).FirstOrDefault();
if (u != null)
{
return (from ux in u.tblInboxes
orderby ux.CreationTS descending
select new UserInboxMsg
{
CreationTS = ux.CreationTS,
ExpirationDate = ux.ExpirationDate,
FromUserId = ux.tblUserId,
HasImage = ux.HasImage,
ImageId = ux.ImageId ?? 0,
IsDeleted = ux.IsDeleted,
IsRead = ux.IsRead,
MsgId = ux.Id,
MsgSize = ux.MessageSize,
ParentId = ux.ParentId,
Title = ux.Title,
ToUserId = userid,
FromUserName = ux.Title,
SenderAvatar = conx.tblMessages.Where(mu=>mu.Id == ux.Id).FirstOrDefault().tblUser.ImageURL,
Message = ux.Message
}).ToList<UserInboxMsg>();
}
else
{
return new List<UserInboxMsg>();
}
}
}

If in the entity-framework, there is a foreign key reference between the two tables you could probably do this:
SenderAvatar = conx.tblMessages.FirstOrDefault( mu=>mu.Id == ux.Id).ImageURL,

Try this.
public List<UserInboxMsg> GetUserInboxMsg(IKASLWSEntities conx, int userid)
{
var u = (from m in conx.tblUsers where m.Id == userid select m).FirstOrDefault();
if (u != null && conx != null)
{
return (from ux in u.tblInboxes
orderby ux.CreationTS descending
select new UserInboxMsg
{
...
...
SenderAvatar = conx.tblMessages.Any(mu=>mu.Id == ux.Id) ? (conx.tblMessages.First(mu=>mu.Id == ux.Id).tblUser != null? conx.tblMessages.First(mu=>mu.Id == ux.Id).tblUser.ImageURL : null) : null,
Message = ux.Message
}).ToList<UserInboxMsg>();
}
else
{
return new List<UserInboxMsg>();
}
}
}
if you are getting null for the Avatar, it is either because there are no entries in tblMessages where mu.Id equals ux.Id or the tblMessage entry is there but the tblUser property is null

There are several problems here.
The first is that the second statement is executed in memory, while it's possible to make the whole query run as SQL:
from u in conx.tblUsers where m.Id == userid
from ux in u.tblInboxes
orderby ux.CreationTS descending
select new UserInboxMsg
{
CreationTS = ux.CreationTS,
ExpirationDate = ux.ExpirationDate,
FromUserId = ux.tblUserId,
HasImage = ux.HasImage,
ImageId = ux.ImageId ?? 0,
IsDeleted = ux.IsDeleted,
IsRead = ux.IsRead,
MsgId = ux.Id,
MsgSize = ux.MessageSize,
ParentId = ux.ParentId,
Title = ux.Title,
ToUserId = userid,
FromUserName = ux.Title,
SenderAvatar = conx.tblMessages.Where(mu => mu.Id == ux.Id)
.FirstOrDefault().tblUser.ImageURL,
Message = ux.Message
}
This has three benefits:
you fetch less data from the database
you get rid of the null reference exception, because SQL doesn't have null references. It just returns null if a record isn't found.
you can return the result of this statement without the if-else.
Second, less important, is that you should use a navigation property like Inbox.Messages in stead of joining (sort of) the inbox and its messages. This makes it less likely that you use the wrong join variables and it condenses your code:
SenderAvatar = ux.Messages.
.FirstOrDefault().User.ImageURL,
Now if there is no avatar, there is no avatar. And there's no null reference exception.
(By the way, you can see that I hate these prefixes in class and property names).

I can only guess this part of your code is wrong : SenderAvatar = conx.tblMessages.Where(mu=>mu.Id == ux.Id).FirstOrDefault().tblUser.ImageURL
I think for example you should use (mu=>mu.UserId == ux.Id) instead of (mu=>mu.Id == ux.Id). In your code, you are comparing "Id" of a table to "Id" of another table which normally in one to many relations is wrong. (only works in one to one relations)
I said I can guess because you didn't mention any information about tblInboxes and tblMessages fields. If you could provide me more information about their structure, I could answer in more detail.
By the way to make your code more clear you can use:
var u = conx.tblUsers.FirstOrDefault(m=>m.Id == userid);
instead of
var u = (from m in conx.tblUsers where m.Id == userid select m).FirstOrDefault();
OR
conx.tblMessages.FirstOrDefault(mu=>mu.Id == ux.Id)
instead of
conx.tblMessages.Where(mu=>mu.Id == ux.Id).FirstOrDefault()

how could i handle this linq query and compare it with and int in an if-statement?

i have the following code:
int selectedcourseId = Convert.ToInt32(c1.Text);
var cid = (from g in re.Sections where g.CourseID == selectedcourseId select g.CourseID);
int selectedinstructorid = Convert.ToInt32(c2.Text);
var iid = (from u in re.Sections where u.InstructorID == selectedinstructorid select u.InstructorID);
i want to compare the two (selectedcourseId) with (cid) and (selectedinstructorid) with (iid) in if-statement such as:
if (selectedcourseId = cid && selectedinstructorid = iid)
{
MessageBox.Show("it already exists");
}
i have tried many things that didnt work our because i have limited knowledge.
thank you very much in advance for any comment or answer

You can change your code as: (but it is meaningless for your situation to check this)
if (selectedcourseId == cid.First() && selectedinstructorid == iid.First())
First of all for checking equality in if statement you must use ==, not =. And the second is the IQueryable<T> allows you to execute a query against a specific data source, but it uses deferred execution. For executing it in your case, you can use First().
But, I suggest that you are just learning how to use LINQ and therefore you have written this code.
I don't know what you are trying to achive. But, if you want to search if there is any result with that ID's, the you must use Any():
var result1 = from g in re.Sections where g.CourseID == selectedcourseId select g.CourseID;
var result2 = from u in re.Sections where u.InstructorID == selectedinstructorid select u.InstructorID;
if(result1.Any() && result2.Any()) { ... }
Or, if you want to find if there is any row which has specified CourseID and InstructorID, then you can call one Any():
if(re.Sections.Any(x => x.CourseID == selectedcourseId && x.InstructorID == selectedinstructorid))
{ ... }

Let me try to find the X from this XY-problem. I guess you want to check if there is already a combination of courseid + instructorid. Then use a single query:
var data = from section in re.Sections
where section.InstructorID == selectedinstructorid
&& section.CourseID == selectedcourseId
select section;
if(data.Any())
{
MessageBox.Show("it already exists");
}

You should not do it in two queries, because the two results might be related to two different rows. This would lead to "false positives" when an instructor handles some section, and a course has some instructors, but the two matches do not belong to the same row:
course instructor
------ ----------
100 10
101 15
102 20
If you are looking for a combination (101, 10) it is not enough to see that 100 is present and 10 is present; you need to check that the two belong to the same row in order to consider it a duplicate.
You can fix this by making a "check presence" query, like this:
var existing = re.Sections
.Any(s => s.InstructorID == selectedinstructorid && s.CourseID == selectedcourseId);
if (existing) {
MessageBox.Show("it already exists");
}

if (selectedcourseId = cid && selectedinstructorid = iid)
this will not work, since single '=' is an assignment, not a comparation (which is '==')
also, you can try to do something like this
var cid = (int)((from g in re.Sections where g.CourseID == selectedcourseId select g.CourseID).FirstOrDefault());
so you select the first or default record from your list and cast it to int

How to check in a linq query for null?

I am trying in a linq query to check if one of the fields is null but i get this error whatever i do.
"Non-static method requires a target."
This is my code :
var users = from s in db.Users
where s.DepartmentId == booking.Item.DepartmentId && s.UserEmail != null
select s;
is any way to go through this error and be able to actually check if UserEmail is null?
p.s : i am using asp.net mvc entity framework.

Be sure that the first 2 letters are uppercase (DBNull.Value).
EDIT:
Try to copy your booking item into a local variable.
var departmentId = booking.Item.DepartmentId;
var users = from s in db.Users
where s.DepartmentId == departmentId && s.UserEmail != null
select s;

var users = from s in db.Users
where s.DepartmentId == booking.Item.DepartmentId && s.UserEmail != DbNull.Value
select s;
You need to compare it to DbNull.Value

LinqToSql ...query without no data

i have this query build with Linq as follow
var query15 = (from g in dc.Fix_Hotel_TTVs
from f in dc.Hotel_Meals_TBLs
where g.CityCode == "KEP" && g.MarketID == "IT" && g.Category == "DEL" && g.HotelCodeID == f.HotelCodeID
select new { Code3 = f.WholeCode == ""? "No Data" : f.WholeCode, Hotel3 = f.HotelName == "" ? "No Data" : f.HotelName }).DefaultIfEmpty();
Well i want show in the DataGrid in case the result of the query is null in the fields Code3 and Hotel3 the string ="NO Data" but i can not find out the solution to have this result ,then do you have any idea how to work out this issue?
Thanks,
Jay

I think hard-coding UI constants in SQL query is not very good idea. Just handle that case in UI. I.e. you should have query in your Data Access layer:
var query = from g in dc.Fix_Hotel_TTVs
join f in dc.Hotel_Meals_TBLs // you can use join
on g.HotelCodeID equals f.HotelCodeID
where g.CityCode == "KEP" &&
g.MarketID == "IT" &&
g.Category == "DEL"
select new FooViewModel { // Consider to create class for this data
Code3 = f.WholeCode,
Hotel3 = f.HotelName
};
Then bind data returned from Data Access layer to DataGrid. Also provide default row with No Data text as described in following questions:
Show “No record found” message on a WPF DataGrid when it's empty
How to Display “No Items found” in a Datagrid

You should supply the default values to the DefaultIfEmpty method:
public class HotelGridViewModel
{
public string Code3{get;set;}
public string Hotel3{get;set;}
}
var query15 = (from g in dc.Fix_Hotel_TTVs
from f in dc.Hotel_Meals_TBLs
where g.CityCode == "KEP" && g.MarketID == "IT" && g.Category == "DEL" && g.HotelCodeID == f.HotelCodeID
select new HotelGridViewModel() { Code3 = f.WholeCode, Hotel3 = f.HotelName
).DefaultIfEmpty(new HotelGridViewModel(){ Code3 = "No Data", Hotel3 = "No Data"});

i think a simple solution...
for example i have a two field name and family:
and my query is:
var query=(from p in behzad.main_table
select p).toArray();
and simple decode your query for example
for(int i=0;i<query.count();i++){
if (query[i].name=="Empty or Null")`
dt[i]["name"]="your message for show in gridview"
when your for complete
simple fill datatable into gridview for example
gridview.datasource(dt);
TNX

Searching for text in a database with Entity Framework

I'm writing a UI that allows a someone to lookup users by their first and/or last name. For example, if you typed in "Mike" for the first name and "Jo" for the last name, it would return "Mike Jones", "Mike Johnson" and "Mike Jobs". I use the following LINQ statement for this search:
var users = (from u in context.TPM_USER
where u.LASTNAME.ToLower().Contains(LastName.ToLower())
&& u.FIRSTNAME.ToLower().Contains(FirstName.ToLower())
select u);
(There may or may not be a better way to do a case-insensitive like clause, but this seems to work)
The problem is if the user types in a first or last name, but then leaves the other field empty. If I type in "Mike" for the first name and leave the Last Name field blank, I want to return all Mikes regardless of their last name. The above query returns no results unless both fields are filled in with at least something.
I tried:
var users = (from u in context.TPM_USER
where (LastName == "" || u.LASTNAME.ToLower().Contains(LastName.ToLower()))
&& (FirstName == "" || u.FIRSTNAME.ToLower().Contains(FirstName.ToLower()))
select u);
However, I still get no results unless both fields are filled out. I've verified under the debugger that LastName == "" is indeed true.
UPDATE:
I did some more debugging and this is actually an Oracle issue. The query being generated is:
--Replaced the field list with * for brevity
SELECT * FROM TPMDBO.TPM_USER "Extent1"
WHERE (('jones' = '') OR ((INSTR(LOWER("Extent1".LASTNAME), LOWER('jones'))) > 0)) AND (('' = '') OR ((INSTR(LOWER("Extent1".FIRSTNAME), LOWER(''))) > 0))
Which at first glance appears to be correct. However, Oracle does not seem to correctly short-circuit the phrase ('' = ''). In fact, if I do:
select * from TPM_USER where '' = ''
I get zero rows. I'm not enough of an Oracle expert to know how this query should be written, but either way it's an Entity Framework dialect bug.

Just add the predicates conditionally:
var users = from u in context.TPM_USER select u;
if (!string.IsNullOrWhiteSpace(FirstName))
users = users.Where(u => u.FIRSTNAME.ToLower().Contains(FirstName.ToLower()));
if (!string.IsNullOrWhiteSpace(LastName))
users = users.Where(u => u.LASTNAME.ToLower().Contains(LastName.ToLower()));
Or only the LASTNAME predicate as conditional one.
Later addition:
An expression like Where(u => u.FIRSTNAME.ToLower()... is better to be avoided. They cause any indexes on FIRSTNAME to be ignored, because the field value is converted first and then compared (see here for more details).
There's a big chance you don't need these lower-case conversions. Check the database collation of the field. If it's case-insensitive (CI), which it probably is, you don't need these conversions.

Are you sure that FirstName and LastName aren't null?
You might try writing it like this instead...
string LowerFirstName = (FirstName + "").ToLower();
string LowerLastName = (LastName + "").ToLower();
var users = (from u in context.TPM_USER
where (LowerLastName == "" || u.LASTNAME.ToLower().Contains(LowerLastName))
&& (LowerFirstName == "" || u.FIRSTNAME.ToLower().Contains(LowerFirstName))
select u);

FYI, if anyone runs into this issue with Oracle, here's a workaround:
var users = (from u in context.TPM_USER
where (LastName == null|| u.LASTNAME.ToLower().Contains(LastName.ToLower()))
&& (FirstName == null || u.FIRSTNAME.ToLower().Contains(FirstName.ToLower()))
select u);
This will get converted to:
'' is null
In SQL, which Oracle interprets as true.

You could simply create a conditional statement around your query:
if (String.IsNullOrWhiteSpace(LastName) && !String.IsNullOrWhiteSpace(FirstName))
{
var users = (from u in context.TPM_USER
where (u.FIRSTNAME.ToLower().Contains(FirstName.ToLower()))
select u);
}
else if (String.IsNullOrWhiteSpace(FirstName) && !String.IsNullOrWhiteSpace(LastName))
{
var users = (from u in context.TPM_USER
where (u.LASTNAME.ToLower().Contains(LastName.ToLower()))
select u);
}

May be you can try checking the length of the search terms to see if it is working in Oracle PL/SQL.
var users = (from u in context.TPM_USER
where ((LastName ?? "").Trim().Length == 0 || u.LASTNAME.ToLower().Contains(LastName.ToLower()))
&& ((FirstName ?? "").Trim().Length == 0 || u.FIRSTNAME.ToLower().Contains(FirstName.ToLower()))
select u);