如何计算Linq左联接右侧没有匹配的记录数？

Question

I am trying to analyse a phone bill, which comprises a load of mobile numbers and associated data. Each number either belongs to somebody in the company, or is unassigned.

I am using this query to match the mobile number to an employee record

var billInfo = from  row in csvTable.AsEnumerable()
               join d in employeeList 
               on row.Field<string>("Phone number") equals d.Mobile into t 
                from rt in t.DefaultIfEmpty()
                    select new
                     {
                       empNo = rt == null ? null : rt.EmpNo,
                       totalRows = csvTable.Rows.Count,
                       unassignedNumbers = t.Where(x => x.EmpNo == null).Count(),
                       name = row.Field<string>("User name"),
                       mobile = row.Field<string>("Phone number")
                      };

I want to get the number of unassigned numbers back in the unassignedNumbers field. (ie If there are 100 rows (one per phone number) in the csvTable datatable, and 70 have an employee record matching in employeeList, I want to return 30 as the value - the number of rows with no matching employee).

The following line:

unassignedNumbers = t.Where(x => x.EmpNo == null).Count,

is where the problem is. It always evaluates to zero, rather than the number of rows with an "empty" employee number (empNo).

Answer 1

Use Count() , not Count . The latter is an ICollection<T> property, while you're dealing with an IEnumerable<T> .

As for the logic of your query, you are running t.Where() on t , which is an inner join. Try

var billInfo = (from  row in csvTable.AsEnumerable()
           join d in employeeList 
           on row.Field<string>("Phone number") equals d.Mobile into t 
           from rt in t.DefaultIfEmpty()
           where rt == null
           select row.Field<string>("Phone number")).Count();

Answer 2

Your query is giving you at least one item for each item in your csvTable . The data you want for unassignedNumbers would be exactly one value for the entire data set. So you'd have to do a completely separate query to do that and then just assign that value directly into the results of your query if desired.

int allUnassignedNumbers = 
    (from  row in csvTable.AsEnumerable()
    join d in employeeList 
    on row.Field<string>("Phone number") equals d.Mobile into t 
    from rt in t.DefaultIfEmpty()
    where rt == null
    select 1).Count();

var billInfo = 
    from  row in csvTable.AsEnumerable()
    join d in employeeList 
    on row.Field<string>("Phone number") equals d.Mobile into t 
    from rt in t.DefaultIfEmpty()
    select new
    {
        empNo = rt == null ? null : rt.EmpNo,
        totalRows = csvTable.Rows.Count,
        unassignedNumbers = allUnassignedNumbers,
        name = row.Field<string>("User name"),
        mobile = row.Field<string>("Phone number")
    };