I have ever tried to use readStream and writeStream, but it throw "Uri
without authority: hdfs:/data/_spark_metadata" exception, which is not seen
in normal read mode.
The parquet path I specified is hdfs:///data
Regard,
Junfeng Chen
On Thu, Mar 8, 2018 at 9:38 AM, naresh Goud
wrote:
> change
change it to readStream instead of read as below
val df = spark
.readStream
.format("kafka")
.option("kafka.bootstrap.servers", "host1:port1,host2:port2")
.option("subscribe", "topic1")
.load()
Check is this helpful
https://github.com/ndulam/KafkaSparkStreams/blob/master/SampleStreamApp