1)Patran can import result files in various formats. The results can be displayed on a mesh that corresponds to the analysis model . The resulting vector or fringe plot can be used to create a Spatial FEM Field. The field can be used to apply LBCS, e.g pressure, to another mesh in the Patran db.
2)Most exact is to use same mesh as CFD - one to one node mapping. But probably impractical. Mapping loads generally means that you are changing the mesh distribution. However you "map" the load the change in mesh requires that you "distribute" or spread the load over an area and this process "dilutes" the original spatial resolution. The change in element normal directions due to different element sizes between the original and new mesh mean that element pressures will give rise to different resolved nodal forces. For "accurate" mapping the overall process should take into account the original "OLOAD" and ensure that the mapped "OLOAD" is the same, otherwise you have changed the total load (forces and moments) on the structure. This is non-trivial to achieve.
3)Start simply, try using pressure mapping and ensuring that the second mesh has sufficient elements to capture the pressure variations of the first. Then the OLOAD differences will hopefully not be significant.
This is a topic that aerospace companies have spent years working up their own internal methodologies and processes around the tools and limitations of software and computers. You can read up on the "SPLINEx" options in the Nastran manual developed for aeroelastic mapping and the Spline6/7 methods introduced in version 2005 for coupling aerodynamic and structural models.