[
https://issues.apache.org/jira/browse/SPARK-11057?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14962085#comment-14962085
]
Narine Kokhlikyan commented on SPARK-11057:
-------------------------------------------
Thank you for you quick response [~rxin]
I have one more question :)
Since my goal is to compute the correlation and covariance for column-pair
combinations and those are independent from each other, I think that it is
better to do it in parallel.
After exploring the APIs in spark I came up with smth like this:
1st sequential example:
let's assume these are my combinations and that for now all my columns are
numerical:
combs
res214: Array[(String, String)] = Array((rating,rating), (rating,income),
(rating,age), (income,rating), (income,income), (income,age), (age,rating),
(age,income), (age,age))
this is how I compute the covariances and it works pefectly.
combs.map(x => peopleDF.stat.cov(x._1, x._2)).foreach(println)
2nd - now I want to compute my covariances in parallel:
val parcombs = sc.parallelize(combs)
parcombs.map(x => peopleDF.stat.cov(x._1, x._2)).foreach(println)
Above example fails with a NullpointerException. I'm new to this, probably I'm
doing something unexpected and if you could point it out me that would be great!
Thanks!
Caused by: java.lang.NullPointerException
at org.apache.spark.sql.DataFrame.schema(DataFrame.scala:290)
at
org.apache.spark.sql.execution.stat.StatFunctions$$anonfun$collectStatisticalData$2.apply(StatFunctions.scala:80)
at
org.apache.spark.sql.execution.stat.StatFunctions$$anonfun$collectStatisticalData$2.apply(StatFunctions.scala:80)
at
scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:244)
> SQL: corr and cov for many columns
> ----------------------------------
>
> Key: SPARK-11057
> URL: https://issues.apache.org/jira/browse/SPARK-11057
> Project: Spark
> Issue Type: New Feature
> Components: SQL
> Reporter: Narine Kokhlikyan
>
> Hi there,
> As we know R has the option to calculate the correlation and covariance for
> all columns of a dataframe or between columns of two dataframes.
> If we look at apache math package we can see that, they have that too.
> http://commons.apache.org/proper/commons-math/apidocs/org/apache/commons/math3/stat/correlation/PearsonsCorrelation.html#computeCorrelationMatrix%28org.apache.commons.math3.linear.RealMatrix%29
> In case we have as input only one DataFrame:
> ------------------------------------------------------
> for correlation:
> cor[i,j] = cor[j,i]
> and for the main diagonal we can have 1s.
> ---------------------
> for covariance:
> cov[i,j] = cov[j,i]
> and for main diagonal: we can compute the variance for that specific column:
> See:
> http://commons.apache.org/proper/commons-math/apidocs/org/apache/commons/math3/stat/correlation/Covariance.html#computeCovarianceMatrix%28org.apache.commons.math3.linear.RealMatrix%29
> Let me know what do you think.
> I'm working on this and will make a pull request soon.
> Thanks,
> Narine
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
