I have a class MyObj and a collection IEnumerable.
Some of the columns are wholly empty (i.e. == NULL) across all rows and therefore I want to create an IEnumerable<> of the members of MyObj which hold a non-null value.
If I could predict the members of MyObj which would be of interest I'd do something like:
var part =
from entry in iList
select new {entry.a, entry.c, entry.s};
...but I don't know which members of MyObj I'm interested in at design time - I only know that at runtime.
How can I construct my list??
Thanks,
Tamim Sadikali.
Your question does not make sense.
You're trying to create a type whose members are only known at runtime.
What would you do with the results?
You would not be able to access any properties of the result objects because they might not exist.
If you want to display the data in a grid, and you don't want to display columns which are entirely null, then you should bind the original collection to the grid, then hide some of the columns in the grid.
Wait for release of VS2010, C# 4.0 with it's 'dynamic' type should solve your problem. (Or maybe help you shoot yourself in the foot).
If you are doing this for UI, better hide columns that contain all nulls. For DataGridView in WinForms it may look like this:
foreach (DataGridViewColumn column in dataGridView.Columns)
if (dataGridView1.Rows.Cast<DataGridViewRow>().All(r => r.Cells[column.Name].Value == null))
column.Visible = false;
Related
There is a table, it is a poco entity generated by entity framework.
class Log
{
int DoneByEmpId;
string DoneByEmpName
}
I am retrieving a list from the data base. I want distinct values based on donebyempid and order by those values empname.
I have tried lot of ways to do it but it is not working
var lstLogUsers = (context.Logs.GroupBy(logList => logList.DoneByEmpId).Select(item => item.First())).ToList(); // it gives error
this one get all the user.
var lstLogUsers = context.Logs.ToList().OrderBy(logList => logList.DoneByEmpName).Distinct();
Can any one suggest how to achieve this.
Can I just point out that you probably have a problem with your data model here? I would imagine you should just have DoneByEmpId here, and a separate table Employee which has EmpId and Name.
I think this is why you are needing to use Distinct/GroupBy (which doesn't really work for this scenario, as you are finding).
I'm not near a compiler, so i can't test it, but...
Use the other version of Distinct(), the one that takes an IEqualityComparer<TSource> argument, and then use OrderBy().
See here for example.
Ok I have seen many questions that based on their text could be something like this but not quite. Say I have something like this
(from r in reports
join u in SECSqlClient.DataContext.GetTable<UserEntity>()
on r.StateUpdateReportUserID equals u.lngUserID
select r).
If reports have a bunch of say reportDTO class and I want to select from a list of that DTO but at the same time set one property to a property in userEntity how would I do that? Basically I want all other fields on the report maintained but set a user name from the user table. (There is a reason this is not done in one big query that gets a list of reports)
What I am looking for is something like Select r).Something(SOME LAMBDA TO SET ONE FIELD TO userEntity property).
There is a dirty way to do this, which is
var repQuery = from r in reports ... select new { r, u };
var reps = repQuery.Select(x => { x.r.Property1 = x.u.Property1; return x.r; };
However, When it comes to functional programming (which Linq is, arguably) I like to adhere to its principles, one of which to prevent side effects in functions. A side effect is a change in state outside the function body, in this case the property value.
On the other hand, this is a valid requirement, so I would either use the ForEach method after converting the query to list (ToList()). Foreach is expected to incur side effects. Or I would write a clearly named extension method on IEnumerable<T> (e.g. DoForAll) that does the same, but in a deferred way. See Why there is no ForEach extension method on IEnumerable?.
I'm using OpenXML to open a spreadsheet and loop through the rows of a spreadsheet. I have a linq query that returns all cells within a row. The linq query was ripped straight from a demo on the MSDN.
IEnumerable<String> textValues =
from cell in row.Descendants<Cell>()
where cell.CellValue != null
select (cell.DataType != null
&& cell.DataType.HasValue
&& cell.DataType == CellValues.SharedString
? sharedString.ChildElements[int.Parse(cell.CellValue.InnerText)].InnerText
: cell.CellValue.InnerText);
The linq query is great at returning all cells that have a value, but it doesn't return cells that don't have a value. This in turn makes it impossible to tell which cell is which. Let me explain a little more. Say for instance we have three columns in our spreadsheet: Name, SSN, and Address. The way this linq query works is it only returns those cells that have a value for a given row. So if there is a row of data that has "John", "", "173 Sycamore" then the linq query only returns "John" and "173 Sycamore" in the enumeration, which in turn makes it impossible for me to know if "173 Sycamore" is the SSN or the Address field.
Let me reiterate here: what I need is for all cells to be returned, and not just cells that contain a value.
I've tried to monkey the linq query in every way that I could think of, but I had no luck whatsoever (ie - removing the where clause isn't the trick). Any help would be appreciated. Thanks!
The OpenXML standard does not define placeholders for cells that don't have data. In other words, it's underlying storage in XML is sparse. You could work round this on one of two ways:
Create a list of all "available" or "possible" cells (probably by using a CROSS JOIN type of operation) then "left" joining to the row.Descendants<Cell>() collection to see if the cell reference has a value
Utilize a 3rd party tool such as ClosedXML or EPPlus as a wrapper around the Excel data and query their interfaces, which are much more developer-friendly.
With ClosedXML:
var wb = new XLWorkbook("YourWorkbook.xlsx");
var ws = wb.Worksheet("YourWorksheetName");
var range = ws.RangeUsed();
foreach(var row in range.Rows())
{
// Do something with the row...
// ...
foreach(var cell in row.Cells())
{
// Now do something with every cell in the row
// ...
}
}
The one way I recommend is to fill in all the null cells with blank data so they will be returned by your linq statement. See this answer for how to do that.
I am new to LINQ queries and to EF too, I usually work with MySQL and I can't guess how to write really simples queries.
I'd like to select all results from a table. So, I used like this:
ZXContainer db = new ZXContainer();
ViewBag.ZXproperties = db.ZXproperties.All();
But I see that I have to write something inside All(---).
Could someone guide me in how could I do that? And if someone has any good link for references too, I thank so much.
All() is an boolean evaluation performed on all of the elements in a collection (though immediately returns false when it reaches an element where the evaluation is false), for example, you want to make sure that all of said ZXproperties have a certain field set as true:
bool isTrue = db.ZXproperties.All(z => z.SomeFieldName == true);
Which will either make isTrue true or false. LINQ is typically lazy-loading, so if you're calling db.ZXproperties directly, you have access to all of the objects as is, but it isn't quite what you're looking for. You can either load all of the objects at the variable assignment with an .ToList():
ViewBag.ZXproperties = db.ZXproperties.ToList();
or you can use the below expression:
ViewBag.ZXproperties = from s in db.ZXproperties
select s;
Which is really no different than saying:
ViewBag.ZXproperties = db.ZXproperties;
The advantage of .ToList() is that if you are wanting to do multiple calls on this ViewBag.ZXproperties, it will only require the initial database call when it is assigning the variable. Alternatively, if you do any form of queryable action on the data, such as .Where(), you'll have another query performed, which is less than ideal if you already have the data to work with.
To select everything, just skip the .All(...), as ZXproperties allready is a collection.
ZXContainer db = new ZXContainer();
ViewBag.ZXproperties = db.ZXproperties;
You might want (or sometimes even need) to call .ToList() on this collection before use...
You don't use All. Just type
ViewBag.ZXproperties = db.ZXproperties;
or
ViewBag.ZXproperties = db.ZXproperties.ToList();
The All method is used to determine if all items of collection match some condition.
If you just want all of the items, you can just use it directly:
ViewBag.ZXproperties = db.ZXproperties;
If you want this evaluated immediately, you can convert it to a list:
ViewBag.ZXproperties = db.ZXproperties.ToList();
This will force it to be pulled across the wire immediately.
You can use this:
var result = db.ZXproperties.ToList();
For more information on linq see 101 linq sample.
All is some checking on all items and argument in it, called lambda expression.
I would like to populate a Treeview.
Here is what I have in DB :
table : Box
BoxID
BoxName
table Book :
BookID
BookName
BoxID (fk Box.BoxID)
table Chapter:
ChapterID
ChapterName
BookID (fk Book.BookID)
As you may know a treeview is made up of treenode object.
A treenode object have a text property and a tag property.
The "text" property is the text that it's display on the screen for this node and the "tag" is an "hidden" value (usually uses to identify a node)
So in my case; the fields ending with ID will be used in the "tag" property and the fields ending with Name will be used in the "text" property
example :
so for a book; I will use the BookID field for the "tag" property and BookName field for the "text" property
note : I use a dbml so I have a Book object, Box object and Chapter object and I use linq to get them from the db.
So my question is; what is the best practice to build this tree?
I have a solution but it's really ugly because it looks like I'm duplicating the code.
The problem is that the values I need to extract for the text and tag properties are identified by differents fields name in the db
So for the book level, I need to get the BookID field to populate the tag property of my node; for the box level, I need to get the BoxID field to populate the tag property , ....
How can I make a kind of generic way to do it ?
I hope I made myself clear enough, don't hesitate to ask me questions :)
Thx in advance
Here is what I have for the moment
I get the list of box with a linq (dbml) request.
List<Box> MyListofBox = getMyListofBox();
Treenode tnBox = null;
Treenode tnBook =null;
foreach(Box b in MyListofBox )
{
tnBox = new TreeNode();
tnBox.tag = b.BoxID;
tnBox.text = b.BoxName;
List<Book> MyListofBook = getMyListofBookByBoxID(b.BoxID)
foreach(Book boo in MyListofBook )
{
tnBook = new TreeNode();
tnBook.tag = boo.BookID;
tnBook.text = boo.BookName;
tnBox.nodes.add(tnBook);
}
mytreeview.nodes.add(tnBox);
}
but of course I don't like this solution...
do you have a better way ?
I would extract the you need from the database in the form of a struct, possibly via the anonnoumous type that has been added to C# together with linq. Then I would populate insert this data into the place in the tree.
From what I get, you are trying to get each property separately, which will not work so well, because then you will have to make a call to the database for each separate property, which is very wasteful.
Addition based on what you have added
I do not believe the code can be more compact - the names you call are similar, but not the same and the way you do it was what I was trying to explain earlier.
You could
Define an key/value interface that both Box and Book implement
Define a delegate that returns a TreeNode and create delegate methods that accept Box and Book
However, I think the code is fine as written. Sometimes you just have to code it and there's little point in further abstracting or optimizing it.
The only issue I see in the code is that you're making a database call in a loop. Whether or not that's a problem depends on the application.