spark-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jörn Franke <jornfra...@gmail.com>
Subject Re: [External] Re: Sorting in Spark on multiple partitions
Date Mon, 04 Jun 2018 17:08:47 GMT
How do you load the data? How do you write it?
I fear without a full source code it will be difficult to troubleshoot the issue.

Which Spark version?

Use case is not yet 100% clear to me. You want to set the row with the oldest/newest date
to true? I would just use top or something similar when processing the data.


> On 4. Jun 2018, at 17:33, Jain, Neha T. <neha.t.jain@accenture.com> wrote:
> 
> Hi Jorn,
>  
> I tried removing userid from my sort clause but still the same issue- data not sorted.
>  
> var newDf = data.repartition(col(userid)).sortWithinPartitions(sid,time)
>  
> I am checking the sorting results  by temporary writing this file to Hive as well as
HDFS. Now, when I see the user wise data it is not sorted.
> Attaching the output file for your reference.
>  
> On the basis of sorting within userid partitions, I want to add a flag which marks first
item in the partition as true other items in that partition as false.
> If my sorting order is disturbed, the flag is wrongly set.
>  
> Please suggest what else could be done to fix this very basic scenario of sorting in
Spark across multiple partitions across multiple nodes.
>  
> Thanks & Regards,
> Neha Jain
>  
> From: Jörn Franke [mailto:jornfranke@gmail.com] 
> Sent: Monday, June 4, 2018 10:48 AM
> To: Sing, Jasbir <jasbir.sing@accenture.com>
> Cc: user@spark.apache.org; Patel, Payal <payal.patel@accenture.com>; Jain, Neha
T. <neha.t.jain@accenture.com>
> Subject: [External] Re: Sorting in Spark on multiple partitions
>  
> You partition by userid, why do you then sort again by userid in the partition? Can you
try to remove userid from the sort? 
>  
> How do you check if the sort is correct or not?
>  
> What is the underlying objective of the sort? Do you have more information on schema
and data?
> 
> On 4. Jun 2018, at 05:47, Sing, Jasbir <jasbir.sing@accenture.com> wrote:
> 
> Hi Team,
>  
> We are currently using Spark 2.2.0 and facing some challenges in sorting of data on multiple
partitions.
> We have tried below approaches:
>  
> Spark SQL approach:
> a.      var query = "select * from data distribute by " + userid + " sort by " + userid
+ ", " + time “
>  
> This query returns correct results in Hive but not in Spark SQL.  
> var newDf = data.repartition(col(userud)).orderBy(userid, time)
> var newDf = data.repartition(col(userid)).sortWithinPartitions(userid,time)
>  
>  
> But none of the above approach is giving correct results for sorting of data.
> Please suggest what could be done for the same.
>  
> Thanks & Regards,
> Neha Jain
>  
> 
> This message is for the designated recipient only and may contain privileged, proprietary,
or otherwise confidential information. If you have received it in error, please notify the
sender immediately and delete the original. Any other use of the e-mail by you is prohibited.
Where allowed by local law, electronic communications with Accenture and its affiliates, including
e-mail and instant messaging (including content), may be scanned by our systems for the purposes
of information security and assessment of internal compliance with Accenture policy. Your
privacy is important to us. Accenture uses your personal data only in compliance with data
protection laws. For further information on how Accenture processes your personal data, please
see our privacy statement at https://www.accenture.com/us-en/privacy-policy. 
> ______________________________________________________________________________________
> 
> www.accenture.com
> <test.csv>

Mime
View raw message