Question

I have a table that records distance from Google/Bing between two valid UK postcodes. As the system is used, this table is added to so that the next distance query becomes fast and I do not have to call web services to retrieve distance(s) online.

Here is the table structure:

    OID                  PostcodeA PostcodeB DistanceMeters                          DistanceMiles
    -------------------- --------- --------- --------------------------------------- ---------------------------------------
    1                    BR60PS    BT248DN   788847                                  490
    2                    BR60PS    CM201JA   64426                                   40
    3                    BR60PS    CM82AP    77640                                   48
    4                    BR60PS    CO123AX   131617                                  82
    5                    BR60PS    CT146EL   119366                                  74
    6                    BR60PS    DA110TA   29247                                   18
    7                    BR60PS    DE216AH   262570                                  163
    8                    BR60PS    DL81AB    397524                                  247
    9                    BR60PS    HG27JE    368802                                  229
    10                   BR60PS    IP121AL   144394                                  90
    11                   BR60PS    IP141AH   144183                                  90
    12                   BR60PS    IP209AH   172259                                  107

Now I have a Scalar UDF that I use through Linq-to-SQL and it is defined as follows:

ALTER FUNCTION [dbo].[GetDistanceFromCache]
(@PostcodeA VARCHAR (MAX), @PostcodeB VARCHAR (MAX))
RETURNS DECIMAL
AS
BEGIN
    DECLARE @FoundDistance AS DECIMAL;
    SELECT @FoundDistance = DistanceMiles
    FROM   AddressInfoRecordedDistance
    WHERE  (PostcodeA = @PostcodeA
                    AND PostcodeB = @PostcodeB)
                 OR (PostcodeB = @PostcodeA
                         AND PostcodeA = @PostcodeB);
    RETURN ISNULL(@FoundDistance, -1);
END

I need to know if there is a any faster SQL statement or c# linq to obtain the result? If I retrieve 800 employees and run them against 50 jobs, the system goes into a pause and if I don’t add the DBContext.GetDistanceFromCache() to the set of selects, the time taken is reduced quite significantly.

And here is the stalling query:

        var query =
                    from locum in DbContext.Locums
                    where
                     locum.IsActive == true &&
                     locum.IsAdminMarkedComplete == true &&
                     locum.IsLocumsExciteBan == false &&
                     locum.IsGPHCBan == false &&
                        filterID1.Contains(locum.OID) == false &&
                        filterID2.Contains(locum.OID) == false
                    select new {
                        LocumID = locum.OID,
                        LocumName = locum.FirstName + " " + locum.LastName,
                        locum.MobileNumber,
                        locum.Email,
                        Gender = locum.Gender ? "Male" : "Female",
                        locum.DateofBirth,
                        LocumType = locum.LocumType.Name,
                        **Distance** = DbContext.GetDistanceFromCache(_Postcode, locum.AddressInfo.Postcode),
                        Address = String.Format("{0} {1} {2} {3}",
                                             locum.AddressInfo.House.Length == 0 ? String.Empty : locum.AddressInfo.House + ", ",
                                             locum.AddressInfo.Street.Length == 0 ? String.Empty : locum.AddressInfo.Street + ", ",
                                             locum.AddressInfo.Area.Length == 0 ? String.Empty : locum.AddressInfo.Area + ", ",
                                             locum.AddressInfo.Postcode ?? String.Empty),
                        Postcode = locum.AddressInfo.Postcode,
                        City = locum.AddressInfo.City.Name,
                        County = locum.AddressInfo.City.County.Name,
                        locum.SystemUserID
                    };

Answer 1

I think your query is slow because the DB has to be called each time the LINQ query does select new { ... }, which is as many times as you have result rows.

I would download the data from the table AddressInfoRecordedDistance before matching the result set to it. I’m thinking something like this:

var query = from locum in DbContext.Locums
            where ...
            select new { ... }; // Don't include GetDistanceFromCache here

var airds = from a in DbContext.AddressInfoRecordedDistance
            select a;

foreach (var q in query)
    q.Distance = GetDistanceFromCache(q.PostcodeA, q.PostcodeB, airds);

And there you have it. Of course, you also need the GetDistanceFromCache method:

// The SQL UDF in C# code here (psuedo code, not tested, just "C blunt")
decimal GetDistanceFromCache(string PostcodeA, string PostcodeB, List<...> table)
{
    return (from t in table
            where
                (t.PostcodeA == PostcodeA && t.PostcodeB == PostcodeB) ||
                (t.PostcodeB == PostcodeA && t.PostcodeA == PostcodeB)
            select t).FirstOrDefault().DistanceMiles;
}

Of course, if you are running all of the above a lot of times, you should cache the variable airds.