Question

I have a list of objects that contain a start/end date. I am trying to either find the current one if a current record exists, or failing that, the most recently expired one.

• Current is defined as a non null start date that is in the past and either a null end date (open ended) or and end date in the future.
• The most recently expired could be defined as having no current time period with a maximum end date. In theory, the time spans denoted by each object should never overlap so I’d expect maximum end date to be sufficient for most recently expired.

I was trying to follow this answer to a similar question, but I can’t seem to find the correct entry when there are other entries in the list that are future dated (ie. have a start/end date where both are in the future).

My attempt is as follows where a is start date, b is end date, c is an indicator for me to use while debugging, and d is used to define a list entry type. I am only interested in type 2.

   [TestMethod]
    public void TestRetrievalOfListElements()
    {

        var baseDate = DateTime.Now;

        var list = new[]
        {
             new { a = new DateTime?(baseDate.AddDays(-90)), 
                   b = new DateTime?(baseDate.AddDays(-60)), c = "LongExpired", d=1 },
             new { a = new DateTime?(baseDate.AddDays(-190)), 
                   b = new DateTime?(baseDate.AddDays(-160)), c = "LongestExpired", d=1 },
             new { a = new DateTime?(baseDate.AddDays(-59)), 
                   b = new DateTime?(baseDate.AddDays(+20)), c = "Current", d=1 },
             new { a = new DateTime?(baseDate.AddDays(-159)), 
                   b = new DateTime?(baseDate.AddDays(-91)), c = "LongerExpired", d=1 },
             new { a = new DateTime?(baseDate.AddDays(-90)), 
                   b = new DateTime?(baseDate.AddDays(-60)), c = "LongExpired", d=2 },
             new { a = new DateTime?(baseDate.AddDays(-190)), 
                   b = new DateTime?(baseDate.AddDays(-160)), c = "LongestExpired", d=2 },
             new { a = new DateTime?(baseDate.AddDays(-59)), 
                   b = new DateTime?(baseDate.AddDays(+20)), c = "Current", d=2 },
             new { a = new DateTime?(baseDate.AddDays(-159)), 
                   b = new DateTime?(baseDate.AddDays(-91)), c = "LongerExpired", d=2 },
             new { a = new DateTime?(baseDate.AddDays(+21)), 
                   b = new DateTime?(baseDate.AddDays(+60)), c = "Future", d=2 },

        }.ToList();

        // The following isn't really right either as it doesn't take into account
        // whether d is of type 1 or 2 either. Not sure how to combine aggregates and
        // other conditions ie MIN and of type 2

        var oldestEntry = (from x in list
                          where x.a == list.Min(d => d.a) &&
                          x.d == 2
                          select x).FirstOrDefault();

        Assert.IsTrue(oldestEntry.a == baseDate.AddDays(-190), "Expected oldest date to be 190 days earlier than today");

        var latestEntryThatIsntInTheFuture = 
                          (from x in list
                          from y in list.Where(z => z.d == 2 && z.a <= baseDate)
                          where x.a == list.Max(d => d.a) &&
                          x.d == 2
                          select x).FirstOrDefault();

        Assert.IsTrue(latestEntryThatIsntInTheFuture.a == baseDate.AddDays(-59), "Expected latest date that isn't in the future to be 59 days earlier than today");



    }

Answer 1

This might be what you’re looking for:

// Find the current which is defined as a non null start date that is in the past 
// and either a null end date (open ended) or and end date in the future. 
var now = DateTime.Now;
var current = list
    .Where(x => x.a.HasValue && x.a.Value < now && (!x.b.HasValue || x.b.Value > now))
    .FirstOrDefault();
if (current == null)
{ 
    //Then find most recently expired 
    current = list.Where(x => x.b.HasValue)
                  .OrderByDescending(x => x.b)
                  .FirstOrDefault();
}

So you don’t need an aggregate in the query in my opinion.