Question

我正在尝试根据spark / python中的首选项对人进行分组。在从原始数据进行多次转换后，我的最终RDD是：

[h1,h3]->[1,4]

现在我转换的RDD应该像

   [((u'90249', u'79727', u'49495'), [u'Collecting Sports Cards (Baseball', u' Basketball', u' Football', u' Hockey)']), 
    ((u'79727', u'12512', u'71917'), [u'Collecting Sports Cards (Baseball', u' Basketball', u' Football', u' Hockey)']), 
    ((u'12512', u'27195', u'49495'), [u'Collecting Sports Cards (Baseball', u' Basketball', u' Football', u' Hockey)']), 
    ((u'90249', u'76176', u'49495'), [u'Collecting Sports Cards (Baseball', u' Basketball', u' Football', u' Hockey)']), 
    ((u'79727', u'27195', u'76176'), [u'Collecting Sports Cards (Baseball', u' Basketball', u' Football', u' Hockey)'])]

我说那个人h1和h3有相同的兴趣。

我怎样才能做到这一点？

我的中级数据如下：

export function createHTTP(url:string, headers?:Headers){
  let injector = ReflectiveInjector.resolveAndCreate([
    myHttp,
    {provide:'defaultUrl', useValue:url},
    {provide:'defaultHeaders', useValue:headers || new Headers()},
    ...HTTP_Providers
  ])
  return injector.get(myHttp)
}

以上是中间转换中产生的RDD。我很惊讶地看到了结果，现在我正面临着梳理最终RDD的问题。

TypeError：不可用类型：'list'

当值等于spark时，组键

0 个答案: