Home / Questions / 334461
Accepted
kev
Asked: 2020-03-05 10:00:01 +0800 CST 2020-03-05 10:00:01 +0800 CST

Join data frames into one in R

  • 772

I want to unite these data frames so that I can then work by making regressions between variables of the different data frames. How can I unite all the data frames into one?

library(dplyr)
library(readxl)
library("tidyverse")
library(caret)
data_p <- read_excel("C:/Users/x/Desktop/regresion.xlsx")
data_t_dic <- read_excel("C:/Users/x/Desktop/temperatura_mensual.xlsx", 
                                  sheet = "dic")
data_t_ene <- read_excel("C:/Users/x/Desktop/temperatura_mensual.xlsx", 
                         sheet = "ene")

data_t_feb <- read_excel("C:/Users/x/desktop/temperatura_mensual.xlsx", 
                         sheet = "feb")

df_p <- data.frame(data_p)
df_t_d <- data.frame(data_t_dic)
df_t_e <- data.frame(data_t_ene)
df_t_f <- data.frame(data_t_feb)
r

2 Answers

  1. Best Answer

    Since you're using tidyverse/dplyryou can take advantage of the verb union_all(), as long as each data.framehas the same structure:

    library("tidyverse")

data_p %>%
  union_all(data_t_dic) %>% 
  union_all(data_t_ene) %>% 
  union_all(data_t_feb) -> data_completa

    In base R it is not complex either, if we userbind()

    data_completa <- rbind(data_p, rbind(data_t_dic, rbind(data_t_ene, data_t_ene)))
    • 2

  2. The ´bind_rows()´ function of ´dplyr´ does exactly that.

    It accepts as argument data.frame names separated by commas, in your case it would be

    bind_rows(data_p, data_t_dic, data_t_ene, data_t_feb)

    An advantage (or disadvantage, depending on the particular use case) is that ´bind_rows()´ will produce a joined data.frame even if the order of the columns is not the same, since the binding is done by names and not by positions . Even when not all df's have the same columns it produces an output by padding with 'NA' in data.frames that don't have a column.

    Another interesting feature is that we can pass it the names or directly a list containing data.frames

    Example:

    lista_df <- list(primero = data.frame(a = 1:5, b = letters[1:5]), 
                 segundo = data.frame(a = 6:10, c = TRUE))
bind_rows(lista_df)

a    b    c
1    a   NA
2    b   NA
3    c   NA
4    d   NA
5    e   NA
6  <NA> TRUE
7  <NA> TRUE
8  <NA> TRUE
9  <NA> TRUE
10 <NA> TRUE

    It has served me a lot working with slightly different databases.

    • 2