How to define a global list in python

Question

I have two methods that should write to the same list.

class MySpider():

    def parse_post(self, response):
        commentList = []
        ...  
        commentList.append(someData)

    def parse_comments(self, response):
        commentList = []
        ...  
        commentList.append(someData)

In this code there are two commentList lists but I need a single list where I can append data. I want to access this list in any method of this class. I tried with

class MySpider():

    commentNum = []

    def parse_post(self, response):
        ...  
        commentList.append(someData)

    def parse_comments(self, response):
        ...  
        commentList.append(someData)

But this gives me an error global name commentList is not defined . Any ideas how to have a single list that can be accessed in all methods in that class?

Answer 1

One way is to simply refer to the variable by its full name ( MySpider.commentList ):

class MySpider(object):

    commentList = []

    def parse_post(self, response):
        ...  
        MySpider.commentList.append(someData)

    def parse_comments(self, response):
        ...  
        MySpider.commentList.append(someData)

This way all instances of MySpider will share the same variable .

If you might have multiple instances of MySpider , and want each instance to have its own commentList , then simply create it in the constructor and refer to it as self.commentList :

class MySpider(object):

    def __init__(self):    
        self.commentList = []

    def parse_post(self, response):
        ...  
        self.commentList.append(someData)

    def parse_comments(self, response):
        ...  
        self.commentList.append(someData)

If both versions would work in your case, I'd suggest using the latter.

Answer 2

Looks like you are using Scrapy. If the list is part of an Item, I usually pass that item to another callback using meta parameter of a Request/Response object .

Answer 3

Just do self.commentList.append(someData)

(Note that normal Python style is to use comment_list and some_data , though.)

Answer 4

Just make it a instance attribute:

class MySpider(object):
    def __init__(self):
        self.comment_list = []
    def parse_post(self, response):
        ...  
        self.comment_list.append(someData)

The instance ( self in python by convention, this in java eg) is explicit in python.

If you initialize your array outside methods (like in your 2nd listing), you make it a class attribute (ie static one), that is "global" to all instances, and it should be referenced using the full name MySpider.comment_list or type(self).comment_list if you want to avoid the class name (eg for inheritance). Due to the lookup of attributes, self.comment_list will also work (if the attribute is not found at instance level, the class is looked for) but the distinction is less obvious ("explicit is better than implicit").