You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@spark.apache.org by Sergey Zhemzhitsky <sz...@gmail.com> on 2017/05/13 13:00:19 UTC
Is GraphX really deprecated?
Hello Spark users,
I just would like to know whether the GraphX component should be considered deprecated and no longer actively maintained
and should not be considered when starting new graph-processing projects on top of Spark in favour of other
graph-processing frameworks?
I'm asking because
1. According to some discussions in GitHub pull requests, there are thoughts that GraphX is not under active development and
can probably be deprecated soon.
https://github.com/apache/spark/pull/15125
2. According to Jira activities GraphX component seems to be not very active and quite a lot of improvements are
resolved as "Won't fix" event with pull requests provided.
https://issues.apache.org/jira/issues/?jql=project%20%3D%20SPARK%20AND%20component%20%3D%20GraphX%20AND%20resolution%20in%20(%22Unresolved%22%2C%20%22Won%27t%20Fix%22%2C%20%22Won%27t%20Do%22%2C%20Later%2C%20%22Not%20A%20Bug%22%2C%20%22Not%20A%20Problem%22)%20ORDER%20BY%20created%20DESC
So, I'm wondering what the community who uses GraphX, and commiters who develop it think regarding this Spark component?
Kind regards,
Sergey
---------------------------------------------------------------------
To unsubscribe e-mail: user-unsubscribe@spark.apache.org
Re: Is GraphX really deprecated?
Posted by Sergey Zhemzhitsky <sz...@gmail.com>.
GraphFrames seems promising but it still has a lot of algorithms, which involve
in one way or another GraphX, or run on top of GraphX according to github
repo (
https://github.com/graphframes/graphframes/tree/master/src/main/scala/org/graphframes/lib),
and in case of RDDs and semistructured data it's not really necessary to
include another library that just will delegate to GraphX, which is still
shipped with Spark as the default graph-processing module.
Also doesn't Pregel-like programming abstraction of GraphX (although it is
on top of RDD joins) seem to be more natural than a number of join steps of
GraphFrames? I believe such an abstraction wouldn't hurt GraphFrames too.
On May 14, 2017 19:07, "Jules Damji" <dm...@comast.net> wrote:
GraphFrames is not part of Spark Core as is Structured Streaming; it's
still open-source and part of Spark packages. But I anticipate that as it
becomes more at parity with all GraphX in algorithms & functionality, it's
not unreasonable to anticipate its inevitable wide adoption and preference.
To get a flavor have a go at it https://databricks.com/blog
/2016/03/03/introducing-graphframes.html
Cheers
Jules
Sent from my iPhone
Pardon the dumb thumb typos :)
On May 13, 2017, at 2:01 PM, Jacek Laskowski <ja...@japila.pl> wrote:
Hi,
I'd like to hear the official statement too.
My take on GraphX and Spark Streaming is that they are long dead projects
with GraphFrames and Structured Streaming taking their place, respectively.
Jacek
On 13 May 2017 3:00 p.m., "Sergey Zhemzhitsky" <sz...@gmail.com> wrote:
> Hello Spark users,
>
> I just would like to know whether the GraphX component should be
> considered deprecated and no longer actively maintained
> and should not be considered when starting new graph-processing projects
> on top of Spark in favour of other
> graph-processing frameworks?
>
> I'm asking because
>
> 1. According to some discussions in GitHub pull requests, there are
> thoughts that GraphX is not under active development and
> can probably be deprecated soon.
>
> https://github.com/apache/spark/pull/15125
>
> 2. According to Jira activities GraphX component seems to be not very
> active and quite a lot of improvements are
> resolved as "Won't fix" event with pull requests provided.
>
> https://issues.apache.org/jira/issues/?jql=project%20%3D%20S
> PARK%20AND%20component%20%3D%20GraphX%20AND%20resolution%
> 20in%20(%22Unresolved%22%2C%20%22Won%27t%20Fix%22%2C%20%22Wo
> n%27t%20Do%22%2C%20Later%2C%20%22Not%20A%20Bug%22%2C%20%
> 22Not%20A%20Problem%22)%20ORDER%20BY%20created%20DESC
>
> So, I'm wondering what the community who uses GraphX, and commiters who
> develop it think regarding this Spark component?
>
> Kind regards,
> Sergey
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe e-mail: user-unsubscribe@spark.apache.org
>
>
Re: Is GraphX really deprecated?
Posted by Jacek Laskowski <ja...@japila.pl>.
Hi,
I'd like to hear the official statement too.
My take on GraphX and Spark Streaming is that they are long dead projects
with GraphFrames and Structured Streaming taking their place, respectively.
Jacek
On 13 May 2017 3:00 p.m., "Sergey Zhemzhitsky" <sz...@gmail.com> wrote:
> Hello Spark users,
>
> I just would like to know whether the GraphX component should be
> considered deprecated and no longer actively maintained
> and should not be considered when starting new graph-processing projects
> on top of Spark in favour of other
> graph-processing frameworks?
>
> I'm asking because
>
> 1. According to some discussions in GitHub pull requests, there are
> thoughts that GraphX is not under active development and
> can probably be deprecated soon.
>
> https://github.com/apache/spark/pull/15125
>
> 2. According to Jira activities GraphX component seems to be not very
> active and quite a lot of improvements are
> resolved as "Won't fix" event with pull requests provided.
>
> https://issues.apache.org/jira/issues/?jql=project%20%
> 3D%20SPARK%20AND%20component%20%3D%20GraphX%20AND%20resolution%20in%20(%
> 22Unresolved%22%2C%20%22Won%27t%20Fix%22%2C%20%22Won%27t%
> 20Do%22%2C%20Later%2C%20%22Not%20A%20Bug%22%2C%20%22Not%20A%20Problem%22)%
> 20ORDER%20BY%20created%20DESC
>
> So, I'm wondering what the community who uses GraphX, and commiters who
> develop it think regarding this Spark component?
>
> Kind regards,
> Sergey
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe e-mail: user-unsubscribe@spark.apache.org
>
>