LINQ Expression to return Property value? - c#

I'm trying to create a generic function to help me select thousands of records using LINQ to SQL from a local list. SQL Server (2005 at least) limits queries to 2100 parameters and I'd like to select more records than that.
Here would be a good example usage:
var some_product_numbers = new int[] { 1,2,3 ... 9999 };
Products.SelectByParameterList(some_product_numbers, p => p.ProductNumber);
Here is my (non-working) implementation:
public static IEnumerable<T> SelectByParameterList<T, PropertyType>(Table<T> items,
IEnumerable<PropertyType> parameterList, Expression<Func<T, PropertyType>> property) where T : class
{
var groups = parameterList
.Select((Parameter, index) =>
new
{
GroupID = index / 2000, //2000 parameters per request
Parameter
}
)
.GroupBy(x => x.GroupID)
.AsEnumerable();
var results = groups
.Select(g => new { Group = g, Parameters = g.Select(x => x.Parameter) } )
.SelectMany(g =>
/* THIS PART FAILS MISERABLY */
items.Where(item => g.Parameters.Contains(property.Compile()(item)))
);
return results;
}
I have seen plenty of examples of building predicates using expressions. In this case I only want to execute the delegate to return the value of the current ProductNumber. Or rather, I want to translate this into the SQL query (it works fine in non-generic form).
I know that compiling the Expression just takes me back to square one (passing in the delegate as Func) but I'm unsure of how to pass a parameter to an "uncompiled" expression.
Thanks for your help!
**** EDIT:** Let me clarify further:
Here is a working example of what I want to generalize:
var local_refill_ids = Refills.Select(r => r.Id).Take(20).ToArray();
var groups = local_refill_ids
.Select((Parameter, index) =>
new
{
GroupID = index / 5, //5 parameters per request
Parameter
}
)
.GroupBy(x => x.GroupID)
.AsEnumerable();
var results = groups
.Select(g => new { Group = g, Parameters = g.Select(x => x.Parameter) } )
.SelectMany(g =>
Refills.Where(r => g.Parameters.Contains(r.Id))
)
.ToArray()
;
Results in this SQL code:
SELECT [t0].[Id], ... [t0].[Version]
FROM [Refill] AS [t0]
WHERE [t0].[Id] IN (#p0, #p1, #p2, #p3, #p4)
... That query 4 more times (20 / 5 = 4)

I've come up with a way to chunk the query into pieces - i.e. you give it 4000 values, so it might do 4 requests of 1000 each; with full Northwind example. Note that this might not work on Entity Framework, due to Expression.Invoke - but is fine on LINQ to SQL:
using System;
using System.Collections.Generic;
using System.Linq;
using System.Linq.Expressions;
using System.Reflection;
namespace ConsoleApplication5 {
/// SAMPLE USAGE
class Program {
static void Main(string[] args) {
// get some ids to play with...
string[] ids;
using(var ctx = new DataClasses1DataContext()) {
ids = ctx.Customers.Select(x => x.CustomerID)
.Take(100).ToArray();
}
// now do our fun select - using a deliberately small
// batch size to prove it...
using (var ctx = new DataClasses1DataContext()) {
ctx.Log = Console.Out;
foreach(var cust in ctx.Customers
.InRange(x => x.CustomerID, 5, ids)) {
Console.WriteLine(cust.CompanyName);
}
}
}
}
/// THIS IS THE INTERESTING BIT
public static class QueryableChunked {
public static IEnumerable<T> InRange<T, TValue>(
this IQueryable<T> source,
Expression<Func<T, TValue>> selector,
int blockSize,
IEnumerable<TValue> values) {
MethodInfo method = null;
foreach(MethodInfo tmp in typeof(Enumerable).GetMethods(
BindingFlags.Public | BindingFlags.Static)) {
if(tmp.Name == "Contains" && tmp.IsGenericMethodDefinition
&& tmp.GetParameters().Length == 2) {
method = tmp.MakeGenericMethod(typeof (TValue));
break;
}
}
if(method==null) throw new InvalidOperationException(
"Unable to locate Contains");
foreach(TValue[] block in values.GetBlocks(blockSize)) {
var row = Expression.Parameter(typeof (T), "row");
var member = Expression.Invoke(selector, row);
var keys = Expression.Constant(block, typeof (TValue[]));
var predicate = Expression.Call(method, keys, member);
var lambda = Expression.Lambda<Func<T,bool>>(
predicate, row);
foreach(T record in source.Where(lambda)) {
yield return record;
}
}
}
public static IEnumerable<T[]> GetBlocks<T>(
this IEnumerable<T> source, int blockSize) {
List<T> list = new List<T>(blockSize);
foreach(T item in source) {
list.Add(item);
if(list.Count == blockSize) {
yield return list.ToArray();
list.Clear();
}
}
if(list.Count > 0) {
yield return list.ToArray();
}
}
}
}

Easiest way to do this: Use LINQKit (Free, non-restrictive license)
Working version of code:
public static IEnumerable<T> SelectByParameterList<T, PropertyType>(this Table<T> items, IEnumerable<PropertyType> parameterList, Expression<Func<T, PropertyType>> propertySelector, int blockSize) where T : class
{
var groups = parameterList
.Select((Parameter, index) =>
new
{
GroupID = index / blockSize, //# of parameters per request
Parameter
}
)
.GroupBy(x => x.GroupID)
.AsEnumerable();
var selector = LinqKit.Linq.Expr(propertySelector);
var results = groups
.Select(g => new { Group = g, Parameters = g.Select(x => x.Parameter) } )
.SelectMany(g =>
/* AsExpandable() extension method requires LinqKit DLL */
items.AsExpandable().Where(item => g.Parameters.Contains(selector.Invoke(item)))
);
return results;
}
Example usage:
Guid[] local_refill_ids = Refills.Select(r => r.Id).Take(20).ToArray();
IEnumerable<Refill> results = Refills.SelectByParameterList(local_refill_ids, r => r.Id, 10); //runs 2 SQL queries with 10 parameters each
Thanks again for all your help!

LINQ-to-SQL still works via standard SQL parameters, so writing a fancy expression isn't going to help. There are 3 common options here:
pack the ids into (for example) csv/tsv; pass down as a varchar(max) and use a udf to split it (at the server) into a table variable; join to the table variable
use a table-valued-parameter in SQL Server 2008
have a table on the server that you could push the ids into (perhaps via SqlBulkCopy) (perhaps with a "session guid" or similar); join to this table
The first is the simplest; getting a "split csv udf" is trivial (just search for it). Drag the udf onto the data-context and consume from there.

Pass IQuerable to the Contains function instead of list or array. please see the below example
var df_handsets = db.DataFeed_Handsets.Where(m => m.LaunchDate != null).
Select(m => m.Name);
var Make = (from m in db.MobilePhones
where (m.IsDeleted != true || m.IsDeleted == null)
&& df_handsets.Contains(m.Name)
orderby m.Make
select new { Value = m.Make, Text = m.Make }).Distinct();
when you pass list or array it is passed in form of parameters and its exceed the counts when the list items count is greater than 2100.

You can create your own QueryProvider
public class QueryProvider : IQueryProvider
{
// Translates LINQ query to SQL.
private readonly Func<IQueryable, DbCommand> _translator;
// Executes the translated SQL and retrieves results.
private readonly Func<Type, string, object[], IEnumerable> _executor;
public QueryProvider(
Func<IQueryable, DbCommand> translator,
Func<Type, string, object[], IEnumerable> executor)
{
this._translator = translator;
this._executor = executor;
}
#region IQueryProvider Members
public IQueryable<TElement> CreateQuery<TElement>(Expression expression)
{
return new Queryable<TElement>(this, expression);
}
public IQueryable CreateQuery(Expression expression)
{
throw new NotImplementedException();
}
public TResult Execute<TResult>(Expression expression)
{
bool isCollection = typeof(TResult).IsGenericType &&
typeof(TResult).GetGenericTypeDefinition() == typeof(IEnumerable<>);
var itemType = isCollection
// TResult is an IEnumerable`1 collection.
? typeof(TResult).GetGenericArguments().Single()
// TResult is not an IEnumerable`1 collection, but a single item.
: typeof(TResult);
var queryable = Activator.CreateInstance(
typeof(Queryable<>).MakeGenericType(itemType), this, expression) as IQueryable;
IEnumerable queryResult;
// Translates LINQ query to SQL.
using (var command = this._translator(queryable))
{
var parameters = command.Parameters.OfType<DbParameter>()
.Select(parameter => parameter)
.ToList();
var query = command.CommandText;
var newParameters = GetNewParameterList(ref query, parameters);
queryResult = _executor(itemType,query,newParameters);
}
return isCollection
? (TResult)queryResult // Returns an IEnumerable`1 collection.
: queryResult.OfType<TResult>()
.SingleOrDefault(); // Returns a single item.
}
public object Execute(Expression expression)
{
throw new NotImplementedException();
}
#endregion
private static object[] GetNewParameterList(ref string query, List<DbParameter> parameters)
{
var newParameters = new List<DbParameter>(parameters);
foreach (var dbParameter in parameters.Where(p => p.DbType == System.Data.DbType.Int32))
{
var name = dbParameter.ParameterName;
var value = dbParameter.Value != null ? dbParameter.Value.ToString() : "NULL";
var pattern = String.Format("{0}[^0-9]", dbParameter.ParameterName);
query = Regex.Replace(query, pattern, match => value + match.Value.Replace(name, ""));
newParameters.Remove(dbParameter);
}
for (var i = 0; i < newParameters.Count; i++)
{
var parameter = newParameters[i];
var oldName = parameter.ParameterName;
var pattern = String.Format("{0}[^0-9]", oldName);
var newName = "#p" + i;
query = Regex.Replace(query, pattern, match => newName + match.Value.Replace(oldName, ""));
}
return newParameters.Select(x => x.Value).ToArray();
}
}
static void Main(string[] args)
{
using (var dc=new DataContext())
{
var provider = new QueryProvider(dc.GetCommand, dc.ExecuteQuery);
var serviceIds = Enumerable.Range(1, 2200).ToArray();
var tasks = new Queryable<Task>(provider, dc.Tasks).Where(x => serviceIds.Contains(x.ServiceId) && x.CreatorId==37 && x.Creator.Name=="12312").ToArray();
}
}

Related

Getting all Data from SQL using LINQ with comma-separated IDs

I do have a string of Empids separated by comma like:
EMpID:"2007,2008,2002,1992,1000,2108,1085
and I need to retrieve the records of all those specified employees using LINQ query.
I tried it with looping but I need to get that in efficient and faster way.
Here goes what i did using looping.
string[] EMpID_str = LeaveDictionary["EMpID"].ToString().Split(',');
for (int i = 0; i < EMpID_str.Length; i++)
{
EMpID = Convert.ToInt32(EMpID_str[i]);
//Linq to get data for each Empid goes here
}
But What I need is to use single LINQ or Lambda query to retrieve the same.Without looping
First convert your ,(comma) separated empId to string array like below:
var empArr = EmpId.split(',');
var employeesResult = emplyeeList.Where(x => empArr.contains(x.EmpId.ToString()));
I hope, it will help someone.
If the Ids that you want to fetch are numbers, not strings, then you should not convert the string to an array of strings, but to a sequence of numbers:
IEnumerable<int> employeeIdsToFetch = LeaveDictionary["EMpID"].ToString()
.Split(',')
.Select(splitText => Int32.Parse(splitText));
To fetch all employees with thees Ids:
var fetchedEmployees = dbContext.Employees
.Where(employee => employeeIdsToFetch.Contains(employee.Id))
.Select(employee => new
{
// Select only the employee properties that you plan to use:
Id = employee.Id,
Name = employee.Name,
...
});
You can use the Expression class to build a Func<int, bool> from your string and use it with the Where methode:
var str = "2,5,8,9,4,6,7";
var para = Expression.Parameter(typeof(int));
var body = str.Split(",")
.Select(s => int.Parse(s))
.Select(i => Expression.Constant(i))
.Select(c => Expression.Equal(para, c))
.Aggregate((a, b) => Expression.Or(a, b));
Func<int, bool> func = Expression.Lambda<Func<int, bool>>(body, para).Compile();
and if you this solution to work with linq to SQL just dont compile the expression at the end and let the linq to SQL engine compile it to an efficent SQL expression.
Instead of the Aggregate Method (which will produce an expression with linear complexity) one could use an divide and conquer approach to fold the values into one value.
For example with this class:
public static class Helper
{
public static T EfficientFold<T>(this List<T> list, Func<T, T, T> func)
{
return EfficientFold(list, 0, list.Count, func);
}
private static T EfficientFold<T>(List<T> list, int lowerbound, int upperbound, Func<T, T, T> func)
{
int diff = upperbound - lowerbound;
var mid = lowerbound + diff / 2;
if (diff < 1)
{
throw new Exception();
}
else if (diff == 1)
{
return list[lowerbound];
}
else
{
var left = EfficientFold(list, lowerbound, mid, func);
var right = EfficientFold(list, mid, upperbound, func);
return func(left, right);
}
}
}
and then we can do
var body = str.Split(",")
.Select(s => int.Parse(s))
.Select(i => Expression.Constant(i))
.Select(c => Expression.Equal(para, c))
.ToList()
.EfficientFold((a, b) => Expression.Or(a, b));
which gives the evaluation a complexity of log(n).

How to get distinct values with corresponding data from IEnumerable

I need to be able to return back only the records that have a unique AccessionNumber with it's corresponding LoginId. So that at the end, the data looks something like:
A1,L1
A2,L1
A3,L2
However, my issue is with this line of code because Distinct() returns a IEnumerable of string and not IEnumerable of string[]. Therefore, compiler complains about string not containing a definition for AccessionNumber and LoginId.
yield return new[] { record.AccessionNumber, record.LoginId };
This is the code that I am trying to execute:
internal static IEnumerable<string[]> GetTestDataForSpecificItemType(ItemTypes itemTypeCode)
{
IEnumerable<StudentAssessmentTestData> data = DataGetter.GetTestData("MyTestData");
data = data.Where(x => x.ItemTypeCode.Trim() == itemTypeCode.ToString());
var z = data.Select(x => x.AccessionNumber).Distinct();
foreach (var record in z)
{
yield return new[] { record.AccessionNumber, record.LoginId };
}
}
That's cause you are selecting only that property AccessionNumber by saying the below
var z = data.Select(x => x.AccessionNumber).Distinct();
You probably want to select entire StudentAssessmentTestData record
data = data.Where(x => x.ItemTypeCode.Trim() == itemTypeCode.ToString()).Distinct();
foreach (var record in data)
{
yield return new[] { record.AccessionNumber, record.LoginId };
}
Instead of using Distinct, use GroupBy. This:
var z = data.Select(x => x.AccessionNumber).Distinct();
foreach (var record in z)
{
yield return new[] { record.AccessionNumber, record.LoginId };
}
should be something like this:
return data.GroupBy(x => x.AccessionNumber)
.Select(r => new { AccessionNumber = r.Key, r.First().LoginId});
The GroupBy() call ensures only unique entries for AccessionNumber and the First() ensures that only the first one LoginId with that AccessionNumber is returned.
This assumes that your data is sorted in a way that if there are multiple logins with the same AccessionNumber, the first login is correct.
If you want to choose distinct values based on a certain property you can do it in several ways.
If it is always the same property you wish to use for comparision, you can override Equals and GetHashCode methods in the StudentAssessmentTestData class, thus allowing the Distinct method to recognize how the classes differ from each other, an example can be found in this question
However, you can also implement a custom IEqualityComparer<T> for your implementation, for example the following version
// Custom comparer taking generic input parameter and a delegate function to do matching
public class CustomComparer<T> : IEqualityComparer<T> {
private readonly Func<T, object> _match;
public CustomComparer(Func<T, object> match) {
_match = match;
}
// tries to match both argument its return values against eachother
public bool Equals(T data1, T data2) {
return object.Equals(_match(data1), _match(data2));
}
// overly simplistic implementation
public int GetHashCode(T data) {
var matchValue = _match(data);
if (matchValue == null) {
return 42.GetHashCode();
}
return matchValue.GetHashCode();
}
}
This class can then be used as an argument for the Distinct function, for example in this way
// compare by access number
var accessComparer = new CustomComparer<StudentTestData>(d => d.AccessionNumber );
// compare by login id
var loginComparer = new CustomComparer<StudentTestData>(d => d.LoginId );
foreach (var d in data.Distinct( accessComparer )) {
Console.WriteLine( "{0}, {1}", d.AccessionNumber, d.LoginId);
}
foreach (var d in data.Distinct( loginComparer )) {
Console.WriteLine( "{0}, {1}", d.AccessionNumber, d.LoginId);
}
A full example you can find in this dotnetfiddle
Add a LinqExtension method DistinctBy as below.
public static class LinqExtensions
{
public static IEnumerable<TSource> DistinctBy<TSource, TKey>(this IEnumerable<TSource> source, Func<TSource, TKey> keySelector)
{
HashSet<TKey> seenKeys = new HashSet<TKey>();
foreach (TSource element in source)
{
if (seenKeys.Add(keySelector(element)))
{
yield return element;
}
}
}
}
Use it in your code like this:
var z = data.DistinctBy(x => x.AccessionNumber);
internal static IEnumerable<string[]> GetTestDataForSpecificItemType(ItemTypes itemTypeCode)
{
IEnumerable<StudentAssessmentTestData> data = DataGetter.GetTestData("MyTestData");
data = data.Where(x => x.ItemTypeCode.Trim() == itemTypeCode.ToString());
var z = data.DistinctBy(x => x.AccessionNumber);
foreach (var record in z)
{
yield return new[] { record.AccessionNumber, record.LoginId };
}
}
This is the code that finally worked:
internal static IEnumerable<string[]> GetTestDataForSpecificItemType(ItemTypes itemTypeCode)
{
var data = DataGetter.GetTestData("MyTestData");
data = data.Where(x => x.ItemTypeCode.Trim() == itemTypeCode.ToString());
var z = data.GroupBy(x => new{x.AccessionNumber})
.Select(x => new StudentAssessmentTestData(){ AccessionNumber = x.Key.AccessionNumber, LoginId = x.FirstOrDefault().LoginId});
foreach (var record in z)
{
yield return new[] { record.AccessionNumber, record.LoginId };
}
}
Returns a sequence that looks like similar to this:
Acc1, Login1
Acc2, Login1
Acc3, Login2
Acc4, Login1
Acc5, Login3
You can try this. It works for me.
IEnumerable<StudentAssessmentTestData> data = DataGetter.GetTestData("MyTestData");
data = data.Where(x => x.ItemTypeCode.Trim() == itemTypeCode.ToString());
var z = data.GroupBy(x => x.AccessionNumber).SelectMany(y => y.Take(1));
foreach (var record in z)
{
yield return new[] { record.AccessionNumber, record.LoginId };
}
I'm not 100% sure what you're asking. You either want (1) only records with a unique AccessionNumber , if two or more records had the same AccessionNumber then don't return them, or (2) only the first record for each AccessionNumber.
Here's both options:
(1)
internal static IEnumerable<string[]> GetTestDataForSpecificItemType(ItemTypes itemTypeCode)
{
return
DataGetter
.GetTestData("MyTestData");
.Where(x => x.ItemTypeCode.Trim() == itemTypeCode.ToString())
.GroupBy(x => x.AccessionNumber)
.Where(x => !x.Skip(1).Any())
.SelectMany(x => x)
.Select(x => new [] { x.AccessionNumber, x.LoginId });
}
(2)
internal static IEnumerable<string[]> GetTestDataForSpecificItemType(ItemTypes itemTypeCode)
{
return
DataGetter
.GetTestData("MyTestData");
.Where(x => x.ItemTypeCode.Trim() == itemTypeCode.ToString())
.GroupBy(x => x.AccessionNumber)
.SelectMany(x => x.Take(1))
.Select(x => new [] { x.AccessionNumber, x.LoginId });
}

Searching a DateTime using a search string in Entity Framework code first

I have an MVC project which outputs a nullable DateTime (DateTime?) as a string to the user for each model item as shown below (my user interface has paging so I only do this computation for a small number of records - in other words, this part is okay):
foreach (DocumentEvent item in Model.items)
#(item?.TimeUtc?.ToString() ?? "N/A")
I want to add a search functionality. I've tried searching as follows, but this is not performant because AsEnumerable materializes my list and I am now in the C# world, enumerating through each record:
using (var context = new ClientEventsContext())
var items = context.Events.AsEnumerable().Where(x => {
(x?.TimeUtc?.ToString() ?? "N/A").Contains(model.search)
});
Instead I want to take advantage of my SQL Server database. How can I build an SQL Server-friendly query for the above code without AsEnumerable that will produce the same results as my current logic?
Here is how you can build and use LINQ to Entities compatible conversion from date to string in M/d/yyyy h:mm:ss tt format. Rather that embedding that monster inside the query, I will use a custom "marker" method and will bind the implementation using ExpressionVisitor. This way you can experiment and change the format if needed (even add some controlling arguments) w/o affecting the readability of the query.
First, the implementation:
public static class EFExtensions
{
public static string ToCustomDateFormat(this DateTime value)
{
// Should never happen
throw new InvalidOperationException();
}
public static IQueryable<T> ApplyCustomDateFormat<T>(this IQueryable<T> source)
{
var expression = new CustomDateFormatBinder().Visit(source.Expression);
if (source.Expression == expression) return source;
return source.Provider.CreateQuery<T>(expression);
}
class CustomDateFormatBinder : ExpressionVisitor
{
protected override Expression VisitMethodCall(MethodCallExpression node)
{
if (node.Method.DeclaringType == typeof(EFExtensions) && node.Method.Name == "ToCustomDateFormat")
{
var date = Visit(node.Arguments[0]);
var year = DatePart(date, v => DbFunctions.Right("0000" + v.Year, 4));
var month = DatePart(date, v => v.Month.ToString());
var day = DatePart(date, v => v.Day.ToString());
var hour = DatePart(date, v => (1 + (v.Hour + 11) % 12).ToString());
var minute = DatePart(date, v => DbFunctions.Right("0" + v.Minute, 2));
var second = DatePart(date, v => DbFunctions.Right("0" + v.Second, 2));
var amPM = DatePart(date, v => v.Hour < 12 ? "AM" : "PM");
var dateSeparator = Expression.Constant("/");
var timeSeparator = Expression.Constant(":");
var space = Expression.Constant(" ");
var result = Expression.Call(
typeof(string).GetMethod("Concat", new Type[] { typeof(string[]) }),
Expression.NewArrayInit(typeof(string),
month, dateSeparator, day, dateSeparator, year, space,
hour, timeSeparator, minute, timeSeparator, second, space, amPM));
return result;
}
return base.VisitMethodCall(node);
}
Expression DatePart(Expression date, Expression<Func<DateTime, string>> part)
{
var parameter = part.Parameters[0];
parameterMap.Add(parameter, date);
var body = Visit(part.Body);
parameterMap.Remove(parameter);
return body;
}
Dictionary<ParameterExpression, Expression> parameterMap = new Dictionary<ParameterExpression, Expression>();
protected override Expression VisitParameter(ParameterExpression node)
{
Expression replacement;
return parameterMap.TryGetValue(node, out replacement) ? replacement : node;
}
}
}
then the usage:
var items = context.Events
.Where(x => x.TimeUtc != null &&
x.TimeUtc.Value.ToCustomDateFormat().Contains(model.search))
.ApplyCustomDateFormat();
I found a solution. This syntax seems to work:
using (var context = new ClientEventsContext())
var items = context.Events.Where(x => {
x.TimeUtc.HasValue && x.TimeUtc.Value.ToString().Contains(model.search)
});
Another solution using SqlFunctions library to convert anything to Raw Query and do the job at SQL level, for example:
using (var context = new ClientEventsContext())
var items = context.Events.Where(x =>
SqlFunctions.PatIndex(model.search,
SqlFunctions.DateName('your pattern here', x?.TimeUtc) ?? "N/A").Value > -1
);

Create Order by programming Linq/Lambda

I have this code :
public void CreateOrdering(string field, string direction)
{
//direction : ASC/DESC
var result = context.MyTable
.Where(x => x.Code > 5)
.OrderBy()
.Skip(10)
.Take(5)
.ToList<MyTable>();
}
I rephrase, I have a method, this method receive as string field name for ordering and the direction ("ASC", "DESC")
I'd like create a Order with the field and the direction received in argument. I have to be able to :
I'd like in this Query be able to do an ascending and descending
Set the ordering field by programming, here Code may be later Id or other ...
The ordering must be done on the SQL Server side not on the list returned
Thanks,
You may use reflection in an extension method which allows for linq syntax:
public static IQueryable<TSource> OrderBy<TSource>(this IQueryable<TSource> source, string field, string direction)
{
string orderByMethod = (direction == "ASC") ? "OrderBy" : (direction == "DESC" ? "OrderByDescending" : null);
if(orderByMethod == null) throw new ArgumentException();
var propertyInfo = typeof (TSource).GetProperty(field);
var entityParam = Expression.Parameter(typeof(TSource), "e");
Expression columnExpr = Expression.Property(entityParam, propertyInfo);
LambdaExpression columnLambda = Expression.Lambda(columnExpr, entityParam);
MethodInfo orderByGeneric = typeof (Queryable).GetMethods().Single(m => m.Name == orderByMethod
&& m.GetParameters().Count() == 2
&& m.GetParameters()[0].ParameterType.GetGenericTypeDefinition() == typeof(IQueryable<>)
&& m.GetParameters()[1].ParameterType.GetGenericTypeDefinition() == typeof(Expression<>));
MethodInfo orderBy = orderByGeneric.MakeGenericMethod(new [] {typeof(TSource), propertyInfo.PropertyType});
return (IQueryable<TSource>) orderBy.Invoke(null, new object[] { source, columnLambda });
}
Sample use:
internal class SomeType
{
public string StringValue { get; set; }
}
IQueryable<SomeType> l = new List<SomeType>
{
new SomeType {StringValue = "bbbbb"},
new SomeType {StringValue = "cccc"},
new SomeType {StringValue = "aaaa"},
new SomeType {StringValue = "eeee"},
}.AsQueryable();
var asc = l.OrderBy("StringValue", "ASC");
var desc = l.OrderBy("StringValue", "DESC");
Or for your example:
context.MyTable
.Where(x => x.Code > 5)
.OrderBy(field, direction)
.Skip(10)
.Take(5)
.ToList<MyTable>();
I may have misunderstood your question, but can't you just do:
Ascending
.OrderBy(x => x.Property)
Descending
.OrderByDescending(x => x.Property)
Update
What you need is Dynamic LINQ. However, what you are trying to do it could get quite complicated. As a simple workaround you could do something like:
var result = context.MyTable
.Where(x => x.Code > 15);
if (direction == "ASC")
{
result = result.OrderBy(field);
}
else
{
result = result.OrderByDescending(field);
}
result = result.Skip(10)
.Take(5)
.ToList<MyTable>();
void Main() {
// Ascending by some other property
CreateOrdering(item => item.SomeProperty, SortDirection.Ascending).Dump("Ascending order for SomeClass.SomeProperty");
// Descending by some other property
CreateOrdering(item => item.SomeProperty, SortDirection.Descending).Dump("Descending order for SomeClass.SomeProperty");
// Ascending by the Code property
CreateOrdering(item => item.Code, SortDirection.Ascending).Dump("Ascending order for SomeClass.Code");
// Descending by the Code property
CreateOrdering(item => item.Code, SortDirection.Descending).Dump("Descending order for SomeClass.Code");
}
// I reccomend not using bare strings, and instead use an enum
public enum SortDirection {
Ascending = 0,
Descending = 1
}
// Define other methods and classes here
public List<SomeClass> CreateOrdering<T>(Expression<Func<SomeClass, T>> field, SortDirection direction) {
// query does not get executed yet, because we have not enumerated it.
var query = context.MyTable
.Where(x => x.Code > 5);
if (direction.Equals(SortDirection.Ascending)) {
query = query.OrderBy (field);
} else {
query = query.OrderByDescending (field);
}
// query gets executed when the call ToList is made.
return query.Skip(10)
.Take(5)
.ToList();
}
public static class context {
private static List<SomeClass> _MyTable = new List<SomeClass>() {
new SomeClass("A", 4), new SomeClass("B", 5), new SomeClass("C", 6),
new SomeClass("D", 7), new SomeClass("E", 8), new SomeClass("F", 9),
new SomeClass("G", 10), new SomeClass("H", 11), new SomeClass("I", 12),
new SomeClass("J", 13), new SomeClass("K", 14), new SomeClass("L", 15),
new SomeClass("M", 16), new SomeClass("N", 17), new SomeClass("O", 18)
};
public static IQueryable<SomeClass> MyTable {
get {
return _MyTable.AsQueryable();
}
}
}
public class SomeClass {
public SomeClass(string property, int code) {
this.SomeProperty = property;
this.Code = code;
}
public string SomeProperty { get; set; }
public int Code { get; set; }
}
normally you would do this:
.OrderBy(x => x.yourField)
or
.OrderByDescending(x => x.yourField)
if you need your field to be dynamic, check this answer
If the field is passed as a string (for instance when using an ObjectDataSource), you can map it using a switch:
var qry = context
.MyTable
.Where(x => x.Code > 5);
switch(orderBy) {
case "MyField": qry = qry.OrderBy(r => r.MyField); break;
case "MyField DESC": qry = qry.OrderByDescending(r => r.MyField); break;
}
// By the way, ToList can infer the generic type if you don't
// want to state it explicity
var result = qry.Skip(10).Take(5).ToList();
The query is not executed before the ToList, and at least with EF it is executed on the SQL Server. I admit the switch is quite a lot of boilerplate, but it did turn out to be quite reliable and fast.

Dynamically filtering a column/property in to an EF 4.1 query using C#

I'm trying to create a generic "search engine" in C# using linq. I have a simple search engine that functions and look like the following.
var query = "joh smi";
var searchTerms = query.Split(new char[] { ' ' });
var numberOfTerms = searchTerms.Length;
var matches = from p in this.context.People
from t in searchTerms
where p.FirstName.Contains(t) ||
p.LastName.Contains(t)
group p by p into g
where g.Count() == numberOfTerms
select g.Key;
I want it to be more generic so I can call it like this:
var matches = Search<Person>(dataset, query, p => p.FirstName, p => p.LastName);
I've gotten as far as the following, but it fails with a "The LINQ expression node type 'Invoke' is not supported in LINQ to Entities." System.NotSupportedException.
static IEnumerable<T> Find<T>(IQueryable<T> items, string query,
params Func<T, string>[] properties)
{
var terms = query.Split(' ');
var numberOfParts = terms.Length;
foreach (var prop in properties)
{
var transformed = items.SelectMany(item => terms,
(item, term) => new { term, item });
// crashes due to this method call
var filtered = transformed.Where(p => prop(p.item).Contains(p.term));
items = filtered.Select(p => p.item);
}
return from i in items
group i by i into g
where g.Count() == numberOfParts
select g.Key;
}
I'm certain it's doable, there just has to be a way to compile i => i.FirstName to an Expression<Func<T, bool>>, but that's where my LINQ expertise ends. Does anyone have any ideas?
You should use a Predicate Builder to construct your Or query, something like:
var predicate = PredicateBuilder.False<T>();
foreach (var prop in properties)
{
Func<T, string> currentProp = prop;
predicate = predicate.Or (p => currentProp(p.item).Contains(p.term));
}
var result = items.Where(predicate );
Look into using a Specification Pattern. Check out this blog. Specifically, look at the spec pattern he developed. This is a similar thought to #Variant where you can build a dynamic specification and pass it to your context or repository.
It turns out the content of the queries just needed to be 'Expanded'. I used a library I found here to expand the expressions. I think that allows Linq to Entities to translate it in to sql. You'll notice Expand gets called over and over again; I think all of them are necessary. It works, anyway. Code to follow:
using System.Linq.Expressions;
public static class SearchEngine<T>
{
class IandT<T>
{
public string Term { get; set; }
public T Item { get; set; }
}
public static IEnumerable<T> Find(
IQueryable<T> items,
string query,
params Expression<Func<T, string>>[] properties)
{
var terms = query.Split(new char[] { ' ' },
StringSplitOptions.RemoveEmptyEntries);
var numberOfParts = terms.Length;
Expression<Func<IandT<T>, bool>> falseCond = a => false;
Func<Expression<Func<IandT<T>, bool>>,
Expression<Func<IandT<T>, bool>>,
Expression<Func<IandT<T>, bool>>> combineOr =
(f, g) => (b) => f.Expand(b) || g.Expand(b);
var criteria = falseCond;
foreach (var prop in properties)
{
var currentprop = prop;
Expression<Func<IandT<T>, bool>> current = c =>
currentprop.Expand(c.Item).IndexOf(c.Term) != -1;
criteria = combineOr(criteria.Expand(), current.Expand());
}
return from p in items.ToExpandable()
from t in terms
where criteria.Expand(new IandT<T> { Item = p, Term = t })
group p by p into g
where g.Count() == numberOfParts
select g.Key;
}
}
It can be called via the following code:
var matches = Search<Person>(dataset, query, p => p.FirstName, p => p.LastName);

Categories

Resources