I have a List<T>
that I want to search not for a given item but for an item satisfying a given condition. Given an item in the list I can test which of 4 conditions is true:
A quick glance at the list functions was not encouraging so I'm wondering if anyone knows off a function I can use?
Edit: this is a local temp list so I known that it will be sorted correctly
Edit: BinarySearch looks almost right but in my case I don't have an item to compare with. I'd use Jon Skeet's solution and ignore one arg, but I'm not sure that I can count on it always being the same arg.
New edit: I'll leave the extra binary searches below, as they'll be useful for others, and here's a final option which is I think what you actually want. Your delegate should return a positive number if the item it's looking for is "less than" the one specified, a negative number if it's "greater than" the one specified, and 0 if it's just right.
public static int BinarySearchForMatch<T>(this IList<T> list,
Func<T,int> comparer)
{
int min = 0;
int max = list.Count-1;
while (min <= max)
{
int mid = (min + max) / 2;
int comparison = comparer(list[mid]);
if (comparison == 0)
{
return mid;
}
if (comparison < 0)
{
min = mid+1;
}
else
{
max = mid-1;
}
}
return ~min;
}
Old edit: I'll leave the original answer below, but here are two other options.
The first takes a projection from the source data to a key type, and specifies the key to find. The comparison itself is just done in the default way for that key type:
public static int BinarySearchBy<TSource,TKey>(this IList<TSource> list,
Func<TSource,TKey> projection, TKey key)
{
int min = 0;
int max = list.Count-1;
while (min <= max)
{
int mid = (min + max) / 2;
TKey midKey = projection(list[mid]);
int comparison = Comparer<TKey>.Default.Compare(midKey, key);
if (comparison == 0)
{
return mid;
}
if (comparison < 0)
{
min = mid+1;
}
else
{
max = mid-1;
}
}
return ~min;
}
The second takes a Func instead, to compare an item from the list with the key we're looking for. The code is almost exactly the same, of course - it's just the comparison which changes:
public static int BinarySearchBy<TSource,TKey>(this IList<TSource> list,
Func<TSource,TKey,int> comparer, TKey key)
{
int min = 0;
int max = list.Count-1;
while (min <= max)
{
int mid = (min + max) / 2;
int comparison = comparer(list[mid], key);
if (comparison == 0)
{
return mid;
}
if (comparison < 0)
{
min = mid+1;
}
else
{
max = mid-1;
}
}
return ~min;
}
These are both untested, but do at least compile :)
Original answer:
You can use List<T>.BinarySearch
with an IComparer<T>
. You don't have to write your own implementation of IComparer<T>
- I've got on in MiscUtil which builds an IComparer<T>
from a Comparison<T>
delegate. That only spots the first three conditions, but the binary search will work out the last one from the rest.
In fact, the code is so short I might as well paste it here (sans comments, admittedly).
public sealed class ComparisonComparer<T> : IComparer<T>
{
readonly Comparison<T> comparison;
public ComparisonComparer(Comparison<T> comparison)
{
if (comparison == null)
{
throw new ArgumentNullException("comparison");
}
this.comparison = comparison;
}
public int Compare(T x, T y)
{
return comparison(x, y);
}
}
So you might do something like:
var comparer = new ComparisonComparer<Person>((p1, p2) => p1.ID.CompareTo(p2.ID));
int index = list.BinarySearch(employee, comparer);
MiscUtil also has a ProjectionComparer
you might be interested in - you just specify a projection, just like in OrderBy
with LINQ - but it really depends on your use case.
Are you sure of those conditions? Normally that works if the list is sorted, in which case perhaps you want to use a SortedList<TKey, TValue>
in the first place.
Either way, assuming C# 3.0 or later you probably want the .Where()
method. Write in your condition as a Lambda and let the framework do the search. It should use a search technique appropriate to the collection.
You might want to implement it in a custom comparator for your object (strangely called comparer in C# docs).
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With