How to Convert a Column from String of json to Array in Apache Spark

Question

The schema of data that read from hdfs is:

root
|-- id: string
|-- ext_json: string

while the data in ext_json is like:

[{'a':'1','b':'2'},{'a':'3','b':'4'}]

now I need to convert the data that schema is as follows:

root
|-- id: string
|-- ext_json: array
|    |-- element: struct
|    |    |-- a: string
|    |    |-- b: string

How to do that?

Spark version is 2.0.1