Search code examples
rapache-sparksparkr

SparkR. How to count distinct values for all columns in a Spark DataFrame?

I am wondering if there is a way to count the number of distinct items in each column of a spark dataframe? That is, given this dataset: 

set.seed(123)
df<- data.frame(ColA=rep(c("dog", "cat", "fish", "shark"), 4), ColB=rnorm(16), ColC=rep(seq(1:8),2))
df

I do this in R to get the counts:

sapply(df, function(x){length(unique(x))} )

> ColA ColB ColC 
   4   16    8

How would I go about doing the same thing for this Spark DataFrame? 

sdf<- SparkR::createDataFrame(df)

Any help is greatly appreciated. Thank you in advance. -nate

Solution

  • This works for me in SparkR:

    exprs = lapply(names(sdf), function(x) alias(countDistinct(sdf[[x]]), x))
# here use do.call to splice the aggregation expressions to agg function
head(do.call(agg, c(x = sdf, exprs)))

#  ColA ColB ColC
#1    4   16    8