In the Supplementary Data Catalogs/Halo Structure files (e.g. 'http://www.tng-project.org/api/TNG300-1/files/halo_structure.99.hdf5') I found many NAN values for the contained variables, (e.g. 'a_form', 'M200c'). I would like to know if I am missing something or if these are the values of the simulation.
A common convention is that catalogs are computed only for some halos, and NaN values indicate other halos (that are not computed).
In this case, the catalog docs state "Restricted to halos with total masses greater than 10^8.8M⊙ (for TNG50), 10^9.8M⊙ (for TNG100), and 10^10.8M⊙ (for TNG300)." - does this match up with the numbers/values you see?
Thanks Dylan. It matches except for c200c where there are nans for halos with masses greater than this value (for about 140000 halos).
Dhayaa Anbajagane
18 Dec
Hi Ángeles!
Thanks for your interest in this catalog :) Indeed as Dylan said above, the NaN values correspond to halos that we did not calculate any properties for. The mass limits described by Dylan above are the general ones we set for all properties, and is set by requiring around ~1500 DMO particles within R200c (in the DMO run). For c200c (which is generally more finnicky to compute) we had set a requirement of at least 2000 DMO particles within R200c. So any halo with c200c = NaN is a halo with less than 2000 particles within R200c.
Note that we only used particles from the central subhalo, so this means at least 2000 particles of the central subhalo must be within R200c. Let me know if this clarifies things or if you have more questions on this!
I realized after your question that this detail was not discussed explicitly in our paper (https://arxiv.org/pdf/2109.02713) so my apologies for that, and thanks for alerting me of this!
Hi Dylan,
In the Supplementary Data Catalogs/Halo Structure files (e.g. 'http://www.tng-project.org/api/TNG300-1/files/halo_structure.99.hdf5') I found many NAN values for the contained variables, (e.g. 'a_form', 'M200c'). I would like to know if I am missing something or if these are the values of the simulation.
Thanks in advance
E.g. in 'http://www.tng-project.org/api/TNG300-1/files/halo_structure.99.hdf5' file, I found 17205178 NAN values in all halo properties (the total data is 17625892)
A common convention is that catalogs are computed only for some halos, and NaN values indicate other halos (that are not computed).
In this case, the catalog docs state "Restricted to halos with total masses greater than 10^8.8M⊙ (for TNG50), 10^9.8M⊙ (for TNG100), and 10^10.8M⊙ (for TNG300)." - does this match up with the numbers/values you see?
Thanks Dylan. It matches except for c200c where there are nans for halos with masses greater than this value (for about 140000 halos).
Hi Ángeles!
Thanks for your interest in this catalog :) Indeed as Dylan said above, the NaN values correspond to halos that we did not calculate any properties for. The mass limits described by Dylan above are the general ones we set for all properties, and is set by requiring around ~1500 DMO particles within R200c (in the DMO run). For c200c (which is generally more finnicky to compute) we had set a requirement of at least 2000 DMO particles within R200c. So any halo with c200c = NaN is a halo with less than 2000 particles within R200c.
Note that we only used particles from the central subhalo, so this means at least 2000 particles of the central subhalo must be within R200c. Let me know if this clarifies things or if you have more questions on this!
I realized after your question that this detail was not discussed explicitly in our paper (https://arxiv.org/pdf/2109.02713) so my apologies for that, and thanks for alerting me of this!