Compare summary statistics for two dataframe with same columns

Question

I have two dataframe A & B with same columns in both except primary key (in the real data I have more than 50+ column like that), Now I want to compare the 'summary' stastics (normal R summary() command) for both the dataframe but for comparison purpose I want to see it next to each other as given in the image attached.

So, please check the image in description to get better details

DATAFRAME DPUT STRUCTURE

structure(list(Pkey = c(1, 2, 3, 4, 5), Phy_marks = c(43, 44,  45,
    46, 47), Math_marks = c(34, 34, 45, 32, 21)), .Names = c("Pkey", 
    "Phy_marks", "Math_marks"), row.names = c(NA, -5L), class =
    "data.frame")

structure(list(Pkey = c(11, 12, 13, 14, 15), Phy_marks = c(43,  44,
    45, 46, 47), Math_marks = c(34, 34, 45, 32, 21)), .Names = c("Pkey",
    "Phy_marks", "Math_marks"), row.names = c(NA, -5L), class =
    "data.frame")

PLEASE HELP!!!

You should provide a minimal reproducible example in our to get better help. Have a look at https://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example — MKR, Mar 10 '18 at 04:30
You can click on `Tick` mark in left of an answer to accept it. — MKR, Mar 12 '18 at 10:15

tigerloveslobsters · Accepted Answer · 2018-03-10T05:12:36.657

5

You can utilize the function I created below to compare the two data sets.

library(dplyr)
compare_them <- function(data1,data2) {
  sum1 <- apply(data1,2,summary) %>% data.frame() 
  sum2 <- apply(data2,2,summary) %>% data.frame() 

  names(sum1) <- paste0(names(sum1),"1")
  names(sum2) <- paste0(names(sum2),"2")

  final <- cbind(sum1,sum2)

  final1 <- t(final) 

  final2 <- final1[order(row.names(final1)), ]

  final_1 <- t(final2) %>% data.frame()
  final_1
}

compare_them(mtcars,mtcars*2) %>% View()

data1 variables will have "1" in the end, data2 variables will have "2" in the end. I used mtcars and mtcars*2 as an example. The end results looks like this.

edited Mar 10 '18 at 05:12

answered Mar 10 '18 at 05:00

tigerloveslobsters

626
5
9

Genius :D Thanks! If there is any simpler way also then do let know – Vaibhav Singh Mar 10 '18 at 05:57
Sorry I new can you tell how to accept the answer, couldnt find it – Vaibhav Singh Mar 12 '18 at 09:28
Hi I get undesired result when doing this something like this A1 A2 B1 B2 C1 C2 Length Class Character Mode Character, getting no values as summary should display please help – Vaibhav Singh Mar 18 '18 at 14:45
1

you can use `dplyr::filter()` to only pick up numeric columns in for data1 and data2. For example,, mtcars %>% filter(is.numeric) – tigerloveslobsters Mar 19 '18 at 03:03

MKR · Answer 2 · 2018-03-10T10:22:43.273

An option is to use summarise.all, dcast, unite and separate to calculate desired stats for each data.frame arrange the same.

Note: The sample data provided by OP has been slightly modified for df_b to have different stats than df_a.

library(tidyverse)
library(reshape2)

df_a %>% mutate(Grp = "A") %>%
bind_rows(mutate(df_b, Grp = "B"))  %>% 
  select(-Pkey) %>% 
  group_by(Grp) %>% {
    inner_join(inner_join(inner_join(summarise_all(.,funs(min,mean,median, max)),
        summarise_all(.,funs(Q1 = quantile), probs = 0.25), by = "Grp"),
        summarise_all(.,funs(Q2 = quantile), probs = 0.50), by = "Grp"),
        summarise_all(.,funs(Q3 = quantile), probs = 0.75), by = "Grp"
  )
} %>% as.data.frame() %>%
  gather(key, val, -Grp) %>%
  separate("key", c("sub", "param"), sep = "_") %>%
  unite("sub", c("sub", "Grp"), sep = "_") %>%
  dcast(param~sub, value.var = "val") %>%
  select_at(vars(param, sort(names(select(.,-param)))))

#   param Math.marks_A Math.marks_B Phy.marks_A Phy.marks_B
#1    max         45.0        100.0          47        99.0
#2   mean         33.2         66.4          45        63.6
#3 median         34.0         80.0          45        60.0
#4    min         21.0         24.0          43        25.0
#5     Q1         32.0         40.0          44        44.0
#6     Q2         34.0         80.0          45        60.0
#7     Q3         34.0         88.0          46        90.0

data

df_a <- structure(list(Pkey = c(1, 2, 3, 4, 5), 
Phy.marks = c(43, 44,  45, 46, 47), Math.marks = c(34, 34, 45, 32, 21)), 
.Names = c("Pkey", "Phy.marks", "Math.marks"), 
row.names = c(NA, -5L), class = "data.frame")

df_b <- structure(list(Pkey = c(11, 12, 13, 14, 15), 
Phy.marks = c(90,  44, 60, 25, 99), 
Math.marks = c(24, 40, 80, 88, 100)), 
.Names = c("Pkey", "Phy.marks", "Math.marks"), 
row.names = c(NA, -5L), class = "data.frame")

Compare summary statistics for two dataframe with same columns

2 Answers2