我的目标是将下面的联接转换为PySpark
merged_df = df1.merge(df2,
left_on=['user','timestamp1'],
right_on=['user','timestamp2'],
how='left')
我这样做了，但这不是理想的结果
merged_df = df1.join(df2,
(df1.user==df2.user) & (df1.timestamp1==df2.timestamp2),
how='left')