One quick thing I've been meaning to blog about LINQ grouping and aggregation, that when you think about it makes sense but might not be immediately evident:
If you are doing a count aggregate on a field that implicitly joined to the key of a grouping join, you will get a count for all the rows, regardless of whether or not the field is NULL.

Consider the following LINQ snippet:

from l in Lawyers
join certInfo in

(
from a in Accounts
group a by a.Certificate into g
select new

{
Certificate = g.Key,
CommentCount = g.Key.CertificateComments.Count(),
TotalFees = g.Sum(a => a.TotalAmount),
TotalTariffs = g.Sum(a => a.TotalTariffAmount)
}

) on l equals certInfo.Certificate.Lawyer

This was a small part of a very large and complicated query we were trying to write in LINQ, and we were surprised to get a CommentCount of 4 rather than 1, the value we were expecting.  If you think about what has to be happening behind the scenes though, it's not that surprising.

The correction is simple:

from l in Lawyers
join certInfo in

(
from a in Accounts
group a by a.Certificate into g
select new

{
Certificate = g.Key,
CommentCount = g.Key.CertificateComments.Count(c => c != null),
TotalFees = g.Sum(a => a.TotalAmount),
TotalTariffs = g.Sum(a => a.TotalTariffAmount)
}

) on l equals certInfo.Certificate.Lawyer