Quick Question regarding the same topic. If construct the data frame *df* in the following way
key | value | topic | offset | partition | timestamp ---------------------------------------------------------------- foo1 | value1 | topic1 | ....... foo2 | value2 | topic2 | ....... what happens when I do df.writeStream().format("kafka").option("topic", "hello_topic").start() ? Will it use the hello_topic specified in the option or the topic from row ? Sure I don't need to specify topic as an option if I have it in a row but I wonder what happens internally if this is the case? Thanks! On Thu, Mar 22, 2018 at 7:48 AM, kant kodali <kanth...@gmail.com> wrote: > Thanks all! > > On Thu, Mar 22, 2018 at 2:08 AM, Jorge Machado <jom...@me.com> wrote: > >> DataFrames are not mutable. >> >> Jorge Machado >> >> >> On 22 Mar 2018, at 10:07, Aakash Basu <aakash.spark....@gmail.com> wrote: >> >> Hey, >> >> I faced the same issue a couple of days back, kindly go through the mail >> chain with "*Multiple Kafka Spark Streaming Dataframe Join query*" as >> subject, TD and Chris has cleared my doubts, it would help you too. >> >> Thanks, >> Aakash. >> >> On Thu, Mar 22, 2018 at 7:50 AM, kant kodali <kanth...@gmail.com> wrote: >> >>> Hi All, >>> >>> Is there a mutable dataframe spark structured streaming 2.3.0? I am >>> currently reading from Kafka and if I cannot parse the messages that I get >>> from Kafka I want to write them to say some "dead_queue" topic. >>> >>> I wonder what is the best way to do this? >>> >>> Thanks! >>> >> >> >> >