[Logical replication] Does the initial table data copy break the transactional replication?

Hi All,

When the publication adds new table, the main apply worker would skip
changes upon that table and launches new sync worker to copy the
existing data of that table, and if possible, the sync worker would
apply changes since the copy start lsn.

Then here comes the question, above logic splits the changes of the
new seen table into other transactions applied by sync worker, so it
breaks the original transaction stream?

Regards,
Jinhua Luo