It's because "toJSON" doesn't support Structured Streaming. The current
implementation will convert the Dataset to an RDD, which is not supported
by streaming queries.
On Sat, Sep 9, 2017 at 4:40 PM, kant kodali <kanth...@gmail.com> wrote:
> yes it is a streaming dataset. so what is the problem with following code?
>
> Dataset<String> ds = dataset.toJSON().map(()->{some function that returns a
> string});
> StreamingQuery query = ds.writeStream().start();
> query.awaitTermination();
>
>
> On Sat, Sep 9, 2017 at 4:20 PM, Felix Cheung <felixcheun...@hotmail.com>
> wrote:
>
>> What is newDS?
>> If it is a Streaming Dataset/DataFrame (since you have writeStream there)
>> then there seems to be an issue preventing toJSON to work.
>>
>> ------------------------------
>> *From:* kant kodali <kanth...@gmail.com>
>> *Sent:* Saturday, September 9, 2017 4:04:33 PM
>> *To:* user @spark
>> *Subject:* Queries with streaming sources must be executed with
>> writeStream.start()
>>
>> Hi All,
>>
>> I have the following code and I am not sure what's wrong with it? I
>> cannot call dataset.toJSON() (which returns a DataSet) ? I am using spark
>> 2.2.0 so I am wondering if there is any work around?
>>
>> Dataset<String> ds = newDS.toJSON().map(()->{some function that returns a
>> string});
>> StreamingQuery query = ds.writeStream().start();
>> query.awaitTermination();
>>
>>
>
