The book isn't wrong about that, your understanding is wrong. The term "key preserved table" doesn't mean that table1 has a primary key, it means that the join from table1 to table2 cannot result in duplicated rows from table1. It is the primary key of table2 that ensures this, because it means that every row from table1 will join to exactly 1 row of table2, so each row of table1 can appear only once in the view. On the other hand, table2 is not key preserved, because the same table2 record can appear many times in the view, once for each table1 record it is referenced by.
The term "key preserved" means that any unique key of table1 that appears in the view will also be a unique key for the view; table2 is not key preserved because its unique key is not a unique key for the view.