Parallel (blended) compression:
You'll see some compressors with a "blend" knob, or read about certain recording engineers using parallel compression, and perhaps you'll wonder "what's the point?" since mixing the compressed signal with the original uncompressed signal essentially undoes some of the effect of the compressor. Fair question!
The idea is that sometimes you may want the thickening or increased sustain that can be gotten with compression, but you may not like the way it flattens your dynamic peaks. Sometimes those peaks are what gives your sound its "feel", its sense of depth, power, or vitality. So blending can be a good compromise. This can be done subtly, or as a huge and interesting effect, depending on your compressor settings.
One method is to use the "aux loop" on a mixer, which is typically mixed in parallel with your dry signal. Or you can use a second mixer channel input instead of the aux loop return, giving the compressed signal its own EQ etc. alongside the dry channel. Another approach is if you have an amp or pedal with a parallel effects loop. Such a loop may be fixed at a 50/50 blend, or it may have a blend percentage knob. Unfortunately you can't assume that every effects loop runs in parallel, if you don't see a blend knob, so read the manual. In either case, mixer loop or amp loop, be sure to choose a compressor that will operate correctly at the signal level your specific device's loop was designed for, instrument level or line level.
The other method is to use a comp that has a blend built in. There are increasing numbers of those, especially in pedal form. Some, like the Barber Tone Press or Guyatone ST2, blend the whole frequency range; while others, like the Seymour Duncan Double Back, have the option of blending in only part of the dry frequency range, like highs or mids. Some rack comps, like the Focusrite Compounder, blend in some uncompressed lows in order to make the bottom end sound bigger.
Be aware that whenever you mix a signal with a processed copy of itself, the two waves will interact with each other, and their differences in phase or polarity can sometimes result in a thin sound with weak lows. Any time you add processing to a signal, its phase will be altered or its polarity may be flipped 180 degrees; but this only matters IF the blended result doesn't sound good. If that happens, you can try a different compressor, or add a device that has a phase switch/knob such as the Xotic X-Blender or the Radial Phazer. With a comp that has blend built in, the circuit designer will generally have addressed any phase problems already, so you don't have to worry about it.
Low frequency notes tend to have bigger amplitude spikes, so the trick is to let them be big enough to sound full and strong, while controlling them enough that they don't cause clipping in your amp or mixer. Compression is all about compromise, and blending is all about compromise; so the bottom line about blended compression is you have to experiment with it and find the right balance of compromises, to find out whether it suits your tastes and does what you want.