manodeep
commented
3 years ago Thanks for using Corrfunc and raising this issue. To be clear, the bins are user-input and are never modified within Corrfunc. I will try to clarify broadly about units in Corrfunc.
For wp (and all theory pair-counters) are units-agnostic, with the assumption that the bins (+ pimax for wp,DDrppi) are in the same units as the positions. So the units could be Mpc, kpc, Mpc/h etc, and as long as the bins (+ pimax) and positions are in the same units, the calculation should be fine.
For DDrppi_mocks (and DDsmu_mocks), the inputs are assumed to be in speed-of-light * redshift and converted to Mpc/h units. However, you can perform the conversion yourself and then set the boolean flag is_comoving_dist=True and the calculation will be performed in your preferred units. The same caveat as theory applies - bins, pimax are both assumed to be in the same units as positions.
Does that clarify the situation?
ulukrasulov
General information
Issue description
When Calculating the projected correlation for mocks from SDSS data using DDrppi_mocks, what units are the bins and wp returned in (Mpc or Mpc/h)?
Following from the previously raised issue and the documenation notes I have asssumed it to be in Mpc/h. However, comparing the calculated wp to other published papers within the same catalogue properties the results of wp from corrfunc match very well only if the units are in Mpc and not Mpc/h
In addition to this when calculating the wp for theoretical catalogues with x,y,z using Corrfunc.theory.wp the clustering between theoretical catalogues and SDSS catalogues only match when inputting units of Mpc into Corrfunc.theory.wp for the theoretical catalogues.
Hence it seems to me that the output of projected clustering for mocks could be in units of Mpc and not Mpc/h? I wanted to double check this issue