Address Object:Best Practices
How can I speed up my address processing without any major restructuring of my program?
There are a couple ways of achieving faster speeds when using our objects. Because our products deal with comparing against a large dataset, it makes sense to streamline lookups of your record by organizing the data before it's sent through our objects.
We recommend organizing your data by zip code, to make the most of data caching. We have seen speeds increase from 71,706 records per hour up to 2,000,000 records per hour just by ordering the data by zip code prior to processing.
You can refer to this page for more details on our speed test: http://www.melissadata.com/ssis1/SSIS-Speed-Scenarios.pdf
Another way of improving performance is to lower the number of data files being initialized and accessed.
Often, we see customers enabling CASS-required data files (DPV, LACSLink, and SuiteLink). If CASS reporting and processing is not a need, then do not initialize the DPV, LACSLink, or SuiteLink data files.
RBDI stands for Residential Business Delivery Indicator and is often used by shipping carriers who would like to identify residential addresses to save them money from these deliveries. If identifying business and residential addresses is not a requirement, then do not initialize the RBDI data files.
SuiteLink and SuiteFinder are used to find missing suite and floor information for addresses and their corresponding company name or contact last name. If appending company suite information is not necessary, then do not initialize the SuiteLink data files. Also, if you don't plan on appending residential suite information, then do not initialize SuiteFinder data files.
Each data file contains at least 1Gb of data, and when all are enabled, can bog down the verification process because of the large number of data lookups.