I don't quite follow the details of that. You have to import the data before you can check to see if that data already exists. I think that you'd have to import the data into a interim table and then only import it into the target table from there if if does not already exist. Instead of performing a find on table B to see if a record exists, you can use a relationship to link records i the interim table to table B if the same data exists in B. Then you can perform a find on the interim field to set up a found set of just those records that do not match to a record in B and then you can import them into the target table.
Other options, some quite simpler are possible, but can only be used in certain situations. An Import matching records might be used in some cases and i other cases a validation rule might work to omit existing data automatically from the report using a unique values validation--possibly on a field that uses an auto-enter calc to combine values from several fields.