CODEWITHSUNDEEP
rdplyrrbind
pudspop
pudspop

Reputation: 97

dplyr returns data frame that won't rbind to comparable data frame

I realise there are alternative ways to get the outcome here, but I'm trying to understand why use of rbind in the following code results in a list, rather than a data frame, despite the input of two apparently identical data frames. It presumably relates to the data frame object returned by dplyr after group_by operation, but how can this be fixed?

The aim is to remove duplicates (on the EventValue1 and EventValue2 columns) where EventCode = X, but keep duplicates for EventCode = Y.

df <- data.frame(EventID = c("1", "2", "3", "4", "5", "6", "7", "8", "9"),
                 EventValue1 = c("A", "A", "B", "C", "D", "E", "E", "F", "F"),
                 EventValue2 = c("AA", "AA", "BB", "CC", "DD", "EE", "FF", "FF", "FF"),
                 EventCode = c("X", "X", "X", "X", "X", "X", "X", "Y", "Y"))

# split df by event code
df.x <- subset(df, EventCode == "X")
df.y <- subset(df, EventCode == "Y") 

# remove duplicates in df.x by EventValue1 and EventValue2 
df.x.2 <- df.x %>% 
  group_by(EventValue1, EventValue2) %>%
  slice(which.min(EventID))

# recombine dfs
df <- rbind(df.x.2, df.y) # this returns a list, should be a data frame


# desired outcome

# EventID EventValue1 EventValue2 EventCode 
# 1       A           AA          X
# 3       B           AA          X
# 4       C           AA          X
# 5       D           AA          X
# 6       E           AA          X
# 7       E           AA          X
# 8       F           FF          Y
# 9       F           FF          Y

Upvotes: 0

Views: 97

Answers (2)

Ronak Shah
Ronak Shah

Reputation: 389135

Since your df.x.2 is grouped by EventValue1 and EventValue2 rbind fails. It works if you ungroup the data 

library(dplyr)
rbind(df.x.2 %>% ungroup(), df.y)

#  EventID EventValue1 EventValue2 EventCode
#* <fct>   <fct>       <fct>       <fct>    
#1 1       A           AA          X        
#2 3       B           BB          X        
#3 4       C           CC          X        
#4 5       D           DD          X        
#5 6       E           EE          X        
#6 7       E           FF          X        
#7 8       F           FF          Y        
#8 9       F           FF          Y

Or use the dplyr specific bind_rows which will still keep the grouping

bind_rows(df.x.2, df.y)

Upvotes: 1

ha-pu
ha-pu

Reputation: 581

Use bind_rows instead of rbind:

df <- bind_rows(df.x.2, df.y)
df

# A tibble: 8 x 4
# Groups:   EventValue1, EventValue2 [7]
  EventID EventValue1 EventValue2 EventCode
  <fct>   <fct>       <fct>       <fct>    
1 1       A           AA          X        
2 3       B           BB          X        
3 4       C           CC          X        
4 5       D           DD          X        
5 6       E           EE          X        
6 7       E           FF          X        
7 8       F           FF          Y        
8 9       F           FF          Y

Upvotes: 0

Related Questions