Overriding start_requests is Scrapy not synchronous - scrapy

Question

I'm trying to override Scrapy's start_requests method, but unsuccessful. I'm already fine to iterate through pages. The problem is that now I have to iterate firstly through cities and than pages.
My code looks like this:
URL = "https://example.com/%s/?page=%d"
starting_number = 1
number_of_pages = 3
cities = [] # there are array of cities
selected_city = "..."
def start_requests(self):
for city in cities:
selected_city = city
print "####################"
print "##### CITY: " + selected_city + " #####"
for i in range(self.page_number, number_of_pages, +1):
print "##### page: " + str(i) + " #####"
yield scrapy.Request(url=(URL % (selected_city, i)), callback = self.parse)
print "####################"
In console I see that when crawler starts working it prints all cities and pages, and than only start requests. Therefore as the result my crawler parses only the first city. They work asynchronously, while I need synchronous.
What is the right way to iterate in my case?
Thanks for any help!

Answer 1

My problem was that I used wrongly global variable selected_city in the remaining code.
I thought that on every iteration it would stop to do parse method, and than continue to next iteration. Therefore I set parameter item['city'] = selected_city in parse method.
Now I just pass parameter city through Request's meta parameter.
Sample code:
def start_requests(self):
requests = []
for city in cities:
for i in range(self.page_number, number_of_pages, +1):
requests.append(scrapy.Request(url=(URL % (city, i)), callback = self.parse, meta = {'city': city}))
return requests
And in parse method retrieving by doing: item['city'] = response.request.meta['city']