Question

The following query runs very slow when running the sql against a SQL Server CE database, I had hoped to translate it to linq for EF. Can anyone advise on this, I haven’t created any indexes on SQL Server CE yet. Table2 has 100,000 rows.

SELECT 
    *
FROM
    Table1 T1a
WHERE
    EXISTS
(
SELECT
    NULL
FROM
    Table2 T2
        JOIN Table1 T1b ON T2.Field1 = T2.Field1
WHERE
    T2.SomeID = 12345 AND
    T1a.SomeString = T1b.SomeString
)
ORDER BY
    T1a.SomeString,
    T1a.AnotherString

Answer 1

Considering ONLY the information given.

To me T2.SomeID would be the most obvious place to start with an index. After that you might look at T2.Field1 because of it’s use in the join (and possibly T1.Field1 depending on the # of rows in T1). The next one is likely going to be T1.SomeString both for it’s comparison usage and because it’s being used in a sort.

Other than T2.SomeID however it kind of depends on the makeup of your data, the cardinality of each value, how many rows are in the tables, how many other queries there are. How many updates vs lookups will be done etc. etc. etc.