I measure 20 oscillations time is 48 seconds. If the least count of that clock is 0.1 second.
What is the error for one(single) oscillation time?

48sec/20 = 2.4 sec/oscillation. The least count of the clock is 0.1 second. Therefore, I think the error in the period is 48±0.1/20 = 2.4 ± 0.005 sec/oscillation.

If I did this wrong, I would appreciate someone telling me. It's been a while since I've done this.

one oscillation time is 2.4 s
so error per single oscillation is ±0.1/48*2.4=±0.0005

This is so wrong still the error is 0.1s. Division does not produce better accurate clock ( if you r correct)