Could also be a LM339 IC on the opto board. Start by unplugging your row and column connectors on the CPU board. Power on the game and go into switch edge test to see if the row is still shorted. If so, problem is on the board, probably U20. If not, problem is probably on the opto board. Unplug the opto board and see if the row short goes away.

No, an opto board LM339 can cause an entire row short rendering all switches in the row to remain closed. In this case unplugging the opto board will restore normal operation, except of course the opto switches. I'm sure the LM339 can fail in other ways that only effect optos, but can for sure short and cause an entire row short.
I just went through this a few months ago on my Medieval Madness. One output pin of the bad LM339 was direct zero ohm short to ground, entire row was shorted, even non opto switches.
Reverse your ohm meter leads from what is normal. Put the positive on ground, then in diode test check the shorted row reading and compare to working rows. On mine, all working rows showed open, shorted row showed .750 because of reading through a diode to the shorted LM339 pin.

Yes, white/orange wire on J3-10 goes into opto board to diode D3, and from there to U1 LM339 pin 2. Most likely U1 is the bad LM339 and has a direct short from pin 2 to ground.