reax/c/bonds and atom_modify

Dear all,

While doing some simple tests I noticed that if I use “fix reax/c/bonds” with “atom_modify map array sort 0 2.0” the connection table gives zeros for ALL nb and id_nb. If I comment “atom_modify” command, I get a reasonable table with nonzero values. I don’t quite understand why should this affecting the connection table. Is there anything with assigning ids to atoms?
Another puzzle for me is coarse bond order cutoffs which is 2.537 in the first case and 0 in the other.Thanks for any comment.

with atom_modify:

Timestep 10