In DS4.1, in the examples below (A & B are University/College Dormitory addresses), the US Address cleanser seems to consistently switch/thrash the output.
For example, the first time through the cleanse if A1 is the input, the process essentially output A2. When A2 is fed through the process, it is cleansed into A1. Behind the scenes we're mapping the PRIMARY_SECONDARY_ADDRESS_BEST_DELIVERY_COMPONENT to an output that in subsequent daily runs is fed into the MULTILINE1 input. Thus, for each daily run, the data thrashes from one value to the other.
| Record | MULTILINE1 (input) | MULTILINE2 (input) | PRIMARY_SECONDARY_ADDRESS_BEST_DELIVERY_COMPONENT (output) |
|---|---|---|---|
| A1 | Pacific Lutheran University | Pflueger Hall | Pflueger Hall |
| A2 | Pflueger Hall | Pacific Lutheran University | Pacific Lutheran University |
| B1 | 5900 W Fake Ave | 222 Simpson Hall | 222 Simpson Hall |
| B2 | 222 Simpson Hall | 5900 W Fake Ave | 5900 W Fake Ave |
| C1 | 555 Georgia Ctr | 3333 S Lumpkin St | 3333 S Lumpkin St |
| C2 | 3333 S Lumpkin St | 555 Georgia Ctr | 555 Georgia Ctr |
We've tweaked various Standardization options to no avail. Note that the thrashing seems to happen in other non-University Dormitory addresses as well.
Any suggestions on how to configure the Cleanser to stop the "thrashing"?