RG-174 cable does not offer very good shielding. This can be a significant problem when you are attempting relatively high levels of attenuation as the cable leakage may provide more signal than the attenuator is allowing. Here is an illustration of coaxial cable leakage from Wikipedia:
There will be losses in the "transport mode" that are due to the RF resistance of the conductors and losses in the dielectric material between the two conductors. This is generally expressed as dB/100 feet or dB/100 meters at a given frequency by the cable manufacturer. The radiating loss is much smaller by comparison and is generally not specified by the manufacturer. This needs to be determined experimentally. Cable TV companies, for example, learned long ago that at least two shields were needed in order to prevent signal leakage that could either be used to pirate cable TV service using a simple antenna pointed at the cable or to not run afoul of regulatory requirements.
Now consider your situation. You wish to attenuate a signal from a higher power source so as to reduce it to a level suitable for a sensitive receiver input. If the coaxial cable leaks only 1/1000 of the power that is traveling through the cable, for example, that is a -30 dB signal source that is radiating from the cable. If you now place a 30 dB attenuator at the end of the cable, the leaked signal and the transported signal are at the nearly same power level. Depending on many other factors, this may or may not affect your measurements.
A general rule for any well equipped RF lab is to use at least double shielded coaxial cables. As @ChrisK8NVH suggested, Times Microwave LMR-100 or LMR-195 would be good choices as they are both double shielded. The nice thing about Times Microwave is that they typically rate their shielding effectiveness - both cables are > 90 dB. These cables are also suitable for SMA connections.
When I am doing sensitive RF work, I will also include several ferrite toroid cores that the cable wraps around in order to reduce common mode current that can also upset sensitive experiments.
Can you get away with RG-174 cable for your experiments? Perhaps. But the results you observe may not be in line with your underlying assumptions.