You can, but you shouldn't. Using backdoors to mutate the data in an RDD is a good way to produce confusing and inconsistent results when, e.g., an RDD's lineage needs to be recomputed or a Task is resubmitted on fetch failure.
On Tue, Dec 29, 2015 at 11:24 AM, ai he <heai0...@gmail.com> wrote: > Same thing. > > Say, your underlying structure is like Array(ArrayBuffer(1, 2), > ArrayBuffer(3, 4)). > > Then you can add/remove data in ArrayBuffers and then the change will > be reflected in the rdd. > > > > On Tue, Dec 29, 2015 at 11:19 AM, salexln <sale...@gmail.com> wrote: > > I see, so in order the RDD to be completely immutable, its content > should be > > immutable as well. > > > > And if the content is not immutable, we can change its content, but > cannot > > add / remove data? > > > > > > > > > > -- > > View this message in context: > http://apache-spark-developers-list.1001551.n3.nabble.com/RDD-Vector-Immutability-issue-tp15827p15841.html > > Sent from the Apache Spark Developers List mailing list archive at > Nabble.com. > > > > --------------------------------------------------------------------- > > To unsubscribe, e-mail: dev-unsubscr...@spark.apache.org > > For additional commands, e-mail: dev-h...@spark.apache.org > > > > > > -- > Best > Ai > > --------------------------------------------------------------------- > To unsubscribe, e-mail: dev-unsubscr...@spark.apache.org > For additional commands, e-mail: dev-h...@spark.apache.org > >
