Hi Stephen,
Could you run this script, and give us the output (perhaps in a paste if it's long)? We're wondering if there are grids with particles in your disk, but without the star-specific fields.
http://paste.yt-project.org/show/2007/ Every grid has no creation_time, even though some of them have NumberOfParticles > 0 (e.g. grid.id=143)
Also, this command will uniquely pull out all the fields in your hdf5 files.
$ h5ls -r *.cpu* | cut -f4 -d\/|cut -f1 -d\ |sort -u Dark_Matter_Density Density Metal_Density Temperature TotalEnergy creation_time dynamical_time metallicity_fraction particle_index particle_mass particle_position_x particle_position_y particle_position_z particle_type particle_velocity_x particle_velocity_y particle_velocity_z x-velocity y-velocity z-velocity I also realized that if I use yt-2.3 instead of yt-2.2 I can get "particle_index" and "particle_mass", but if I try "particle_type" I get the same error than with "creation_time". Thanks! Fernando.