Linq distinct not working correctly

Question

I'm having a strange problem with a linq query. I'm using LINQPad 4 to make some a query that uses regular expression using LinqToSQL as the LinqPad driver.

Here's the query that I'm trying to make :

(from match in
from s in SystemErrors
select Regex.Match(s.Description, "...")
select new 
{
  FamilyCode = match.Groups["FamilyCode"].Value,
  ProductPrefix = match.Groups["ProductPrefix"].Value,
  BillingGroup = match.Groups["BillingGroup"].Value,
  Debtor = match.Groups["Debtor"].Value
}).Distinct()

As you can see I'm trying to extract data from a text description in a log table using groups. The query works, but the Distinct doesn't want to work, it returns a line for all Match.

I have read that distinct should work with anonymous type, matching each property. Even more strange is that distinct does actually do something, it orders the values alphabetically by the FamilyCode (and then by ProductPrefix, etc.).

Has someone an idea on why this isn't working? Thanks

Here is what is displayed in the SQL tab of LinqPad :

DECLARE @p0 NVarChar(1000) = 'Big Regexp'
DECLARE @p1 NVarChar(1000) = 'FamilyCode'
DECLARE @p2 NVarChar(1000) = 'ProductPrefix'
DECLARE @p3 NVarChar(1000) = 'BillingGroup'
DECLARE @p4 NVarChar(1000) = 'Debtor'

SELECT DISTINCT [t2].[Description] AS [input], [t2].[value], [t2].[value2], [t2].[value3], [t2].[value4], [t2].[value5]
FROM (
    SELECT [t1].[Description], [t1].[value], @p1 AS [value2], @p2 AS [value3], @p3 AS [value4], @p4 AS [value5]
    FROM (
        SELECT [t0].[Description], @p0 AS [value]
        FROM [SystemError] AS [t0]
        ) AS [t1]
    ) AS [t2]

Answer 1

var result = from eachError in SystemErrors
             let match = Regex.Match(eachError.Description, "...")
             group eachError by new 
             {
              FamilyCode = match.Groups["FamilyCode"].Value,
              ProductPrefix = match.Groups["ProductPrefix"].Value,
              BillingGroup = match.Groups["BillingGroup"].Value,
              Debtor = match.Groups["Debtor"].Value
             }
             into unique
             select unique.key;

When you use Distinct(), it's distinct by pointer to each object, not value because select new {} is object type not value type. Try using group by instead.

Answer 2

On the other hand, you can use .Distinct(IEqualityComparer<T>) overload and provided EqualityComparer for the object that you want to process.