CODEWITHSUNDEEP
rdplyrgroup-summaries
Juanchi
Juanchi

Reputation: 1166

summarize data with two functions in dplyr

Considering this example dataframe:

d <- read.table(text="
  trt rep y  
  1   1   30   
  1   1   50   
  1   1   70   
  1   2   0   
  1   2   0   
  1   2   0   
  2   1   10   
  2   1   0   
  2   1   0   
  2   2   5   
  2   2   0   
  2   2   .   
  "
  , header = TRUE, check.names = F, na.strings = ".")

I'm trying to obtain a summary table with two operations of the "y" variable.

The first new column should have the simple mean values by trt for every rep: 

by_rep1 = d %>% 
  group_by(trt, rep) %>%
  summarise(sev = mean(na.omit(y)))

and the second one, the proportion of positives values by trt for every rep. 

by_rep2 = d %>% 
  group_by(trt, rep) %>%
  summarise_each(funs(round(mean(.>0, na.rm=TRUE),2)), y)

I'm doing this long, because I'm don't have idea how to do it in one step:

inner_join(by_rep1, by_rep2, by = c("trt", "rep"))  

#    trt   rep    mean_y     y
#  (int) (int)     (dbl) (dbl)
#1     1     1 50.000000  1.00
#2     1     2  0.000000  0.00
#3     2     1  3.333333  0.33
#4     2     2  2.500000  0.50

does someone knows how to do that in a single step, joining both functions?

Upvotes: 0

Views: 2258

Answers (2)

akrun
akrun

Reputation: 887851

We can also use data.table

library(data.table)
setDT(d)[, .(mean_y = mean(y, na.rm = TRUE), y = round(mean(y > 0,
               na.rm = TRUE), 2)) , .(trt, rep)]
#   trt rep    mean_y    y
#1:   1   1 50.000000 1.00
#2:   1   2  0.000000 0.00
#3:   2   1  3.333333 0.33
#4:   2   2  2.500000 0.50

Or this can be done using only base R

do.call(data.frame, aggregate(y~., d, FUN = function(x) 
     c(mean_y=mean(x, na.rm=TRUE), y=round(mean(x > 0, na.rm=TRUE),2)), na.action=NULL))

Upvotes: 1

akuiper
akuiper

Reputation: 215117

You can put them into a single summarize statement:

d %>% group_by(trt, rep) %>% summarise(mean_y = mean(y, na.rm = T), 
                                       y = round(mean(y > 0, na.rm = T), 2))
Source: local data frame [4 x 4]
Groups: trt [?]

    trt   rep    mean_y     y
  (int) (int)     (dbl) (dbl)
1     1     1 50.000000  1.00
2     1     2  0.000000  0.00
3     2     1  3.333333  0.33
4     2     2  2.500000  0.50

Upvotes: 4

Related Questions