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)
.