PipelinedRDD“对象没有属性”\u jdf pyspark数据帧

Question

我试图在pyspark中构建一个电影推荐模型，当我尝试训练该模型时，我一直得到“PipelinedRDD”对象没有属性“\u jdf”

导入库：

from pyspark.ml.recommendation import ALS

加载数据：

df = sc.textFile('ml-latest-small/ratings.txt').map(lambda x: x.split("::").toDF(["user", "item","rating","timestamp"]))

X_train, X_test = df.randomSplit([0.8, 0.2])

als = ALS(rank=5, maxIter=10, seed=0)
model = als.fit(X_train)

错误：

AttributeError: 'PipelinedRDD' object has no attribute '_jdf'