Если в каждой строке одинаковое количество результатов, вы можете сделать это:
def parse(self, response):
addresses = []
areas = []
prices_stable = []
prices_drop = []
prices_decreased = []
links = []
for i in response.css('span.dir'):
addresses.append(i.css('b::text').extract())
for l in response.css('div.datos'):
areas.append(l.css('i::text').extract())
for x in response.css('div.opciones'):
prices_stable.append(x.css('span.eur::text').extract())
for o in response.css('div.opciones'):
prices_drop.append(o.css('div.mp_pvpant.baja::text').extract())
for y in response.css('div.opciones'):
prices_decreased.append(y.css('span.eur_m::text').extract())
for u in response.css('div.datos'):
links.append(u.css('a::attr(href)').extract_first())
for address, area, price_stable, price_drop, price_decreased, link in zip(addresses, areas, prices_stable, prices_drop, prices_decreased, links):
yield {
'address': address,
'area': area,
'price_stable': price_stable,
'price_drop': price_drop,
'price_decreased': price_decreased,
'link': link,
}