This is an old Idea for Design which I've used in modified form. This circuit is a good first start with both a series and shunt switch but with the switch outside the feedback loop.
As John and I have both pointed out elsewhere, the switch can be placed inside the feedback loop to virtually eliminate delta-Rds distortion. It's not shown that way here:
Removing the Constraints of C-MOS Bilateral Switches
Removing the Constraints of C-MOS Bilateral Switches (pdf 587K)
There is a better method with the FET or CMOS switch placed inside the feedback loop. This is a discrete FET version inside the loop used in the dBx 3Bx.
There is no shunt in this version, only the series switch. Note that the op-amp input resistor tied to Q4 is off-page.
Ignore the CR19/CR30 network which is mis-drawn. (Looks like the anode of CR19 should go to the bottom of R43 10M.)
dBx 3Bx FET Switch Detail
This is one I hastily drew. At the time I called it the "spoon fed edition" because the person I was drawing it for was being a real snarky PITA with no imagination. (It was not on this forum.)
The sections can be repeated to form a selector. Each input can have it's own gain set independent of the other inputs. IF DG413 switches are used, the inverters can be eliminated.
The delta-Rds of the switches are now in series with the inverting input. Since the inverting input is very high impedance, THD from delta-Rds is virtually eliminated.
CMOS Switch Inside the Loop
