Closed jotelha closed 3 years ago
ncjoin
operates under the assumption that id
is only necessary if the atoms are not sorted. ncjoin
sorts the atoms, therefore you no longer require id
. The id
is then identical to the index of the atom.
This is of course only true if there are no gaps in the ids, which may happen if you remove atoms at some point during the simulation.
I see. I have introduced those modifications, https://github.com/IMTEK-Simulation/code-snippets/pull/4. The default behavior of excluding id
won't change.
Sounds good...
Fixed with #4
@pastewka is there are particular reason for always excluding the "id" field from joint trajectories by default,
https://github.com/IMTEK-Simulation/code-snippets/blob/5d8e6a2f2f4e29d43db9d99150683de2f382ffd8/imteksimcs/cli/NetCDF/ncjoin.py#L128 ?
That renders the atoms unidentifiable after joining, I would like to remove that line, do you object?
As an example, the structure of a single netcdf file generated by LAMMPS might look like this,
and the structure of multiple such files concatenated via
ncjoin.py
like this,with the only important difference at the
id
field,I want them both to have the
id
field.