Hi Armando,
This one is slightly tricky - because among the best practice for EDMs one should avoid the following: blank cells, cells or rows with duplicate data:
Remove Blank Columns and Duplicate Rows from the Data Source File
Thus, the trick for you is to try a method whereby the Source Data File (aka, the original CSV from which the EDM is indexed) doesn't include that type of data. I realize this is not straightforward, because the "Account Number" as such seems to be the key identifier for an account, so if an account (but not all) can have multiple email addresses associated, it would not necessarily be better to try different column/row format:
e.g., having each Account Number in ONE ROW only, but having multiple columns for "extra" email addresses - because for accounts with only one contact, the extra email columns would be blank.
What I wonder if whether EDM is fully the best method for doing this. What about authoring a policy that has your EDM data which is unique:
Account Number, Company
And pairing that with a condition using a Reusable Recipient Pattern - that includes ALL recipient emails?
I know that isn't as fully accurate, because the EDM data might create False Positives when detected alongside a recipient in the list but from a different company, but that might be less common than the issue you are reporting now.
Hope this proves helpful - again, it might take some other permutations of what you are after to get best results.
Thanks,
------------------------------
Stephen Heider
DLP Sr Prn Info Writer | Symantec Enterprise Division | DLP Support
Broadcom
stephen.heider@broadcom.com | broadcom.com
------------------------------