Spark - Scala：当json数据分布在多行时，读取json文件为dataframe不起作用？

Question

JSON数据：

{ "blogID" : "FJY26J1333", "date" : "2012-04-01", 
"name" : "vpxnksu", "comment" : "good stuff"} 
{"blogID" : "VSAUMDFGSD", "date" : "2012-04-12", "name" : "yhftrcx", "comment" : "another comment"} 

代码：

val dataFrame=sqlContext.read.json("sample.json") 
dataFrame.show() 

输出：

_corrupt_record  blogID  comment   date  name 
{ "blogID" : "FJY... null  null    null  null 
"name" : "vpxnksu... null  null    null  null 
null     VSAUMDFGSD another comment 2012-04-12 yhftrcx 

我怎样才能把它读作两个记录？

Answer 1

确保每行其一JSON对象，这样在源文件中：

{ "blogID" : "FJY26J1333", "date" : "2012-04-01", "name" : "vpxnksu", "comment" : "good stuff"} 
{ "blogID" : "VSAUMDFGSD", "date" : "2012-04-12", "name" : "yhftrcx", "comment" : "another comment"}