Group by

By far the most common query I've every run in SQL Server is:

Select SomeColumn, Count(*) as [Count]
from SomeTable
group by SomeColumn
order by 2 desc

Here's the equivalent in Linq:

var tableSummary =
someTable.GroupBy(p=> p.SomeColumn).Select(group => new {
	Count = group.Count(),
	SomeColumn = group.Key
}).OrderByDescending(x => x.Count);
tableSummary.Dump();

Here's a more complex example:

var data = new Person[] {
	new Person { Name = "Jim", Height = 1000},
	new Person { Name = "Jack", Height = 2000},
	new Person { Name = "Jack", Height = 2300},
	new Person { Name = "Jim", Height = 5000},
	new Person { Name = "Jim", Height = 12}
};
var groups = data.GroupBy(p => p.Name, (key, values)
	=> new {
		Name = key,
		Count = values.Count(),
		Avg_Height = values.Average(a => a.Height),
		Total_Height = values.Sum(a => a.Height),
		Name_Length = values.Average(a => a.Name.Length),
		Name_Length2 = key.Length,
		Max_Height = values.Max(a => a.Height),
		Min_Height = values.Min(a => a.Height),
		Variance_Height = values.Max(a => a.Height) - values.Min(a => a.Height),
		Members = values}
		).Where(g => g.Count > 1).OrderByDescending(gg => gg.Variance_Height);
groups.Dump();

result is 1 row per name with many aggregated properties regarding that name and child rows: of type person, with name matching their parent group

(Supporting code....

public class Person
{
	public string Name { get; set; }
	public int Height { get; set; }
}

...)

Group by

See also