I want to create a dataframe with summary features for all columns in the dataframe using summarize_all() function and the dataframe has NA values.
Input->
a b
0 1
Output->
a.min a.max a.mean a.sd b.min b.max b.mean b.sd
0 0 0 0 1 1 1 1
Code:
df<- df%>%
summarize_all(funs( min , max ,mean, sd))
How to handle null values in this code?
In the dataframe posted in question, there are NA values because there is not enough data to compute standard deviation. See below:
> df1<-data.frame("a"=0,"b"=1)
> df1%>%
+ summarize_all(funs( min , max ,mean, sd))
a_min b_min a_max b_max a_mean b_mean a_sd b_sd
1 0 1 0 1 0 1 NaN NaN
> df2<-data.frame("a"=c(0,1,2,3),"b"=c(1,3,5,7))
> df2%>%
+ summarize_all(funs( min , max ,mean, sd))
a_min b_min a_max b_max a_mean b_mean a_sd b_sd
1 0 1 3 7 1.5 4 1.290994 2.581989
In case you have NA values in the dataset, using na.rm=T will solve your purpose:
> df3<-data.frame("a"=c(0,1,NA,3),"b"=c(1,3,5,7))
# with na.rm=T
> df3%>%
+ summarize_all(funs( min , max ,mean, sd),na.rm=T)
a_min b_min a_max b_max a_mean b_mean a_sd b_sd
1 0 1 3 7 1.333333 4 1.527525 2.581989
# without na.rm=T
> df3%>%
+ summarize_all(funs( min , max ,mean, sd))
a_min b_min a_max b_max a_mean b_mean a_sd b_sd
1 NA 1 NA 7 NA 4 NaN 2.581989
