scrapy的name变量_scrapy 跨组件传递参数

引用 @玉帛 的方法，可以解决问题，实现“从MongoDB读取start_url，对start_url进行处理，生成特征值，再将特征值传递给pipeline作为collection表名”的操作，具体解决方案如下。

Spider中：

def start_requests(self):

client = pymongo.MongoClient('localhost',27017)

db_name = 'Sina'

db = client[db_name]

collection_set01 = db['UrlsQueue']

datas=list(collection_set01.find({},{'_id':0,'url':1,'status':1}))

for data in datas:

if data.get('status') == 'pending':

url=data.get('url')

pattern='(?<=/)([0-9a-zA-Z]{9})(?=\?)'

if re.search(pattern,url):

collection_name=re.search(pattern,url).group(0)

start_url='https://weibo.cn/comment/'+collection_name+'?ckAll=1'

collection_set01.update({'url':url},{'$set':{'status':'proccessing'}})

break

else:

pass

client.close()

yield Request(url=start_url,callback=self.parse, cookies=cookie, meta={'collection_name':collection_name})

从数据库中获取start_url，提取特征值，并对其处理，带meta参数发送request

def parse(self,response):

collection_name=response.meta['collection_name']

......

for i in range(0,len(node)):

item['collection_name']=collection_name

yield item

parse()从response中解析数据的同时提取回传的meta参数

Pipeline中：

def close_spider(self, spider):

self.db['UrlsQueue'].update({'status':'proccessing'},{'$set':{'status':'finished'}})

self.client.close()

def process_item(self, item, spider):

self.collection_name=item.pop('collection_name')

self.db[self.collection_name].insert_one(dict(item))

return item

pop掉collection_name参数即可

非常感谢 @玉帛 的帮助