Cross-sectional sampling allows you to sample records throughout the file, from beginning to end. You can specify cross-sectional sampling for the input name-and-address file using a FILEDF parameter. With the NTHSEL parameter, you can specify cross-sectional for the input name-and-address file only.
To determine the number to enter in the FILEDFs 7-character DECIMAL FRACTION field, divide the number of records you want to process by the number of records in the file. This gives you a decimal number. Drop the decimal, and type the first 7 digits of the number. For example, suppose you wanted to process 2,000 records from a file containing 102,986 records. To select 2,000 records for processing, divide 102,986 into 2,000. The answer is .0194201153. Drop the decimal and keep 7 digits (0194201) to type in the DECIMAL FRACTION field. This ensures that the 2,000 records processed are evenly distributed throughout the file. You can use this sampling with the NTHSEL parameter as well.
Example - Skipping and Limiting Records