FWIW I once met a cable that was intended to be re-used, I can't recall exactly what it was, 12 core 2.5 possibly, that tested generally OK but a couple of cores had a few megohms to earth. Turns out someone had put a core drill through about 8 of the cores, which had then been peeled away from the cable and the breaks bridged out with 3" lengths of wire with a choc-block at each end. But the whole lot was in contact with damp concrete and the blocks were corroded and turning green and leaking to earth, and one had melted and the insulation burned away along the cable. The surprising thing is that the continuity was fine at first inspection.
My point here (and Davesparks') is that the fault you detected on the cable may not in itself be a problem in your application but it is an indicator of an unknown degree of damage. It could be a small screw through the cable, it could be seriously mangled, you don't know. But you do know for certain that it is damaged.
The correct procedure for testing insulation is to connect all cores and the armour together to start with. Disconnect one core, test it to the remainder, then reconnect and move to the next core. 20 tests required (don't forget the armour to all the cores) and each result shows as near as possible the quality of insulation of that core alone. Any fault will show up exactly twice.
What testing device are you using? Don't forget that a voltage reading on a supposedly disconnected core is totally meaningless! If your meter has a high input impedance, capacitive coupling can raise the voltage very close to supply voltage, especially as in 19-core armoured it's possible for one core to be completely surrounded by cores at line potential throughout its length. Even with an undamaged cable I would not be surprised to see an unused core float up to 230V on my meter, although of course it won't light a test lamp.
