VMD join structures not working

Dear lammps experts:
In vmd I used the tcl command “pbc join …” to join structures that have long bonds due to the periodic box condition. But after I use this command, there are still bonds within a molecule that span the whole box.
I get the reason is that my molecule is too long. (a polymer chain containing 18 monomer is a molecule). If this is the reason, how can I join the bonds spanning the whole box?

Thank you for checking

This is a question about a VMD plugin, not LAMMPS, so it should be posted on the vmd-l mailing list and not here.

Besides, why would you even need to use pbc join on a LAMMPS trajectory when you can output unwrapped coordinates directly?

Thank you Axel,
(Yes it is a vmd issue. ) It sounds very helpful and interesting that the trajectory can be outputted unwrapped. If I used the unwrap trajectory, I think the system will be very sparse and keep growing as the simulation runs. (maybe close to this post: [Outputting LAMMPS trajectory file - #2 by akohlmey]).
I want to have a box of the system, but any bonds that span the whole box will be joined. Does the unwrapped output trajectory is what I want in this situation?

Thank you for checking,