CODEWITHSUNDEEP
rdataframeduplicates
WillBell
WillBell

Reputation: 45

Remove duplicate rows in one column based on another column and keep other columns intact

I have tried a lot of solutions found here and none have seamed to work correctly; unique function got me closest. My data looks like:

id   second   var1   var2
100   20       3      4
100   21       3      3
100   22       4      3
100   23       4      3
100   24       4      4 
100   22       3      3
100   23       3      3

It repeats about ten seconds usually every 300 or so seconds. Each session is around 1200 seconds. I would like delete duplicate seconds within a session and take the mean of whatever is being collapsed in var1 and var 2 or if not the mean keeping either original value is OK. Everything I have tried only removes duplicates if var1 and var2 are non unique?

Upvotes: 2

Views: 1018

Answers (2)

M--
M--

Reputation: 28825

You can do it using base-R functions:

aggregate(. ~ id+second, dat, mean)

##    id second var1 var2
## 1 100     20  3.0    4
## 2 100     21  3.0    3
## 3 100     22  3.5    3
## 4 100     23  3.5    3
## 5 100     24  4.0    4

Using dplyr package it can be done (similar to previous answer with a little of tweaking):

library(dplyr)
dat %>% group_by(id, second) %>% 
        summarise_all(mean)

## # A tibble: 5 x 4
## # Groups:   id [?]
##      id second  var1  var2
##   <int>  <int> <dbl> <dbl>
## 1   100     20   3.0     4
## 2   100     21   3.0     3
## 3   100     22   3.5     3
## 4   100     23   3.5     3
## 5   100     24   4.0     4

If you want to remove duplicates based on two columns (keys), data.table package would be a good option:

library(data.table)
unique(setDT(dat), by = c('id','second'))

#     id second var1 var2
# 1: 100     20    3    4
# 2: 100     21    3    3
# 3: 100     22    4    3
# 4: 100     23    4    3
# 5: 100     24    4    4

Data:

 dat <- structure(list(id = c(100L, 100L, 100L, 100L, 100L, 100L, 100L                       
     ), second = c(20L, 21L, 22L, 23L, 24L, 22L, 23L), var1 = c(3L,                          
     3L, 4L, 4L, 4L, 3L, 3L), var2 = c(4L, 3L, 3L, 3L, 4L, 3L, 3L)), .Names = c("id",        
     "second", "var1", "var2"), .internal.selfref = <pointer: 0x02af24a0>, row.names = c(NA, 
     7L), class = c("data.table", "data.frame"))

Upvotes: 0

Preston
Preston

Reputation: 8177

This will create a new dataframe with the requirements that you asked for.

To explain, you don't actually need to delete anything, you just need to group the val1/2s by the common values, in this case id and second.

library(tidyverse)

new_df <- df %>%
  group_by(id, second) %>%
  summarise(var1 = mean(var1),
            var2 = mean(var2)
            )

Upvotes: 1

Related Questions