Sergey,
once again, many thanks for taking the time to give such detailed help.
John.

>Hi John,
>
>You don't have to concatenate fields. The second one should be faster.
>

>select k_prempl, k_prdedmst
>	from c_PenDetail pd
>	WHERE NOT EXISTS ( SELECT * FROM prdedemp
>		WHERE k_prdedmst = pd.k_prdedmst AND k_prempl = pd.k_prempl)
>
>* or
>select pd.k_prempl, pd.k_prdedmst
>	from c_PenDetail pd
>	LEFT JOIN prdedemp pr ON pr.k_prdedmst = pd.k_prdedmst AND pr.k_prempl = pd.k_prempl
>	WHERE pr.k_prdedmst IS NULL
>
>

>>here is the revised code that does work as I want. I have a cross reference table that matches employees with deductions using keys K_PREMPL and K_PRDEDEMP. I need to add new records where K_PREMPL+K_PRDEDEMP exist in the cursor c_PenDetail but do not exist in prdedemp.
>>

* Add any missing deduction types for each employee.
>>TEXT TO mSqlCommand noshow pretext 15 TEXTMERGE
>>  insert into <<xfprdedemp>>
>>  (k_prempl, k_prdedmst)
>>
>>  select k_prempl, k_prdedmst
>>  from c_PenDetail
>>
>>  where transform(k_prdedmst,'9999999999')+ transform(k_prempl,'9999999999')
>>  not in (select transform(k_prdedmst,'9999999999')+ transform(k_prempl,'9999999999') from prdedemp)
>>ENDTEXT
>>&mSqlCommand

>>
>>Is there a better way than using the transform to concatenate the fields?