Creating value-added halo catalogs through cross-matching

All halo catalogs come with an integer ID column providing a unique identifier of the (sub)halo in the catalog. This tutorial demonstrates two different examples of how you can use the crossmatch function to exploit this column to create “value-added” versions of your halo catalogs. In Example 1, we’ll show how to combine information from two partially overlapping halo catalogs. In Example 2, we’ll show how to create new columns for a subhalo catalog storing the properties of the host halo, e.g., host mass \(M_{\rm vir}^{\rm host}\).

For a closely related tutorial, see Cross-matching galaxy and halo catalogs.

Example 1: Combining information from different halo catalogs

When analyzing halo catalogs, it’s a common situation for you to have two different versions of a halo catalog, one with halo properties that you wish to transfer to the other. In general, the two versions may only partially overlap, as different cuts may have have been applied to the catalogs. We’ll demonstrate this scenario using the FakeSim halo catalog that is randomly generated on-the-fly, but the same calculation applies equally well to real halo catalogs, or generally any structured data table with an object ID.

>>> from halotools.sim_manager import FakeSim
>>> halocat1 = FakeSim()
>>> halo_table1 = halocat1.halo_table

>>> halocat2 = FakeSim()
>>> mask = halocat2.halo_table['halo_mvir'] > 1e11
>>> halo_table2 = halocat2.halo_table[mask]

Now let’s add some new column information to halo_table2 and use the crossmatch function to transfer this information to halo_table1. This function returns the indices providing the correspondence between the rows in halo_table1 that have matches in halo_table2.

>>> import numpy as np
>>> halo_table2['some_new_column'] = np.random.random(len(halo_table2))

The halo catalog column halo_id is a Long giving a unique identifier to every halo and subhalo in the halo catalog, so we can use that column to match one object to the other.

>>> from halotools.utils import crossmatch
>>> halo_table1['transferred_column'] = np.zeros(len(halo_table1), dtype = halo_table2['some_new_column'].dtype)
>>> idx_table1, idx_table2 = crossmatch(halo_table1['halo_id'], halo_table2['halo_id'])
>>> halo_table1['transferred_column'][idx_table1] = halo_table2['some_new_column'][idx_table2]

Now for those objects in halo_table1 that are also in halo_table2, the values from the some_new_column column will be stored in the transferred_column; rows without a matching entry will still be set to their initial value of zero.

Example 2: Transferring host halo properties to their subhalos

When analyzing catalogs that include subhalos, one very common kind of calculation that is done over and over is to group subhalos according to some property of the host halo, such as host halo mass. Such calculations become easy when there is a column in your data table storing the associated host halo property, and in order to create such a column, you need to cross-match the halo_id column against the halo_hostid column. As described in Rockstar halo and subhalo nomenclature conventions, for the case of subhalos, the halo_hostid column points to the halo_id of the host halo. So we use the crossmatch function to add new columns to the halo catalog such that some property of the host halo is transferred onto all of its subhalos.

>>> halocat = FakeSim()
>>> t = halocat.halo_table
>>> idx_table1, idx_table2 = crossmatch(t['halo_hostid'], t['halo_id'])
>>> t['host_halo_mvir'] = t['halo_mvir'] # initialize the new column
>>> t['host_halo_mvir'][idx_table1] = t[idx_table2]['halo_mvir']