GENERALLY BLACK HAS BETTER PAWN FORMATION IN SICILIAN, SO HE HAS ADVANTAGE IN ENDINGS..
How is black ending better in the sicilian defense ?
OK - my ideas are these: First the pawns on d6&e6 often means the black king is safe on e7. Since the black king is often closer to the center by delaying castling, often this is a small plus. The white queenside majority is often hard to get moving with a knight on c3 and the king on c1. It's even harder when black has pressure on the c file. If the black rooks have pressure on c3 a black knight on c4 can put a lot of pressure on the black queenside. If white throws his center pawns forward to f5 or e5 it creates weaknesses in the center that gives blacks king a path forward or out posts. OTOH, if black throws pawns forward on queenside it is a standard minority attack and can open targets for his/her rooks. Often enough white gives up the bishop pair to save a few tempos for attack. Another small plus. If white overextends for attack, blaclk wins the endgame. Otherwise I thing black is often a bit more comfortable in the endgame. -Bill

People say that white has space and an attack but black has compensation in a better ending.
Why is that ?
People who "Are they beginners?" Based on what evidence? Do these people know statistics?
It is totally wrong.
Statistics of Master games on OTB is "White 55% score, black 45%"
Statistics of ICCF games , the scores range from "51%-49%" to "53%-47%" depending on rating level.
As chess is a known draw game if both side doesnt blunder, it should be 50%-50% . But due to the time limit, the first move advantage slightly favour white and white usually get >50% scores in large data.
"

This is a Sicilian where black won. Every move was forced for both sides.
Doesn't white win after 18.Bxg6?
White's attack seems unstoppable after 18...Rg8 19.Bxf7+ Kxf7 20.Qh3.
Yes - I'd say it's probably the clearest and most useful book out there on the subject it covers. Reading it in the first half of 2013 exactly correlates with my rating rising from ~125 ECF (1638 FIDE) to ~155 ECF (1830 FIDE) - I like to think that that's more than just coincidence.