Question

我们说我有三张表重叠。

    A   B   C   D
A   12  16  17  14
B   62  66  9   85
C   37  31  59  75
D   74  76  89  25

    A   B   E   F
A   12  16  11  19
B   62  66  57  28
E   24  21  4   51
F   7   1   68  22

    C   D   E   F
C   59  75  77  80
D   89  25  88  30
E   67  87  4   51
F   39  69  68  22

我想按行和按列组合它们，没有任何重复的行或列，并且行和列名称保持不变。

    A   B   C   D   E   F
A   12  16  17  14  11  19
B   62  66  9   85  57  28
C   37  31  59  75  77  80
D   74  76  89  25  88  30
E   24  21  67  87  4   51
F   7   1   39  69  68  22

三天后，我设法将这一点拼凑起来（在here，here，here的帮助下，以及可能已经忘记的其他人的帮助下：

#Import tables as dataframes
file.names <- dir(pattern = ".tab")
for(i in 1:length(file.names)){
  nam <- paste("table.", i, sep = "")  #rename the data as table.1 ... table.n
  assign(nam, as.data.frame(as.matrix(read.delim(file.names[i],
         row.names=1, header=TRUE, sep="\t", stringsAsFactors=FALSE))))
}

#Import an empty file (i.e. just column and row names) 
#that you will fill with your smaller data tables
out.file <- as.data.frame(as.matrix(read.delim("Blank_table.csv",
                                               row.names=1, header=TRUE, sep=",")))

#Create a list of the dataframes
file.names = lapply(ls(pattern = "table.[0-9]"), get)

#Add columns that we can use for merging
#because using 'merge' on dataframes destroys row names
out.file$rows <- rownames(out.file)
for(i in 1:length(file.names)){
  rownams <- rownames(file.names[[i]])
  file.names[i] <- lapply(file.names[i], cbind, rows = rownams)
}

#Combine the tables
for(i in 1:length(file.names)){
  file <- file.names[i]
  out.file <- aggregate(. ~ rows, data = merge(out.file, file, all = TRUE),
                        na.action = na.pass, FUN = mean, na.rm = TRUE)
}

这就是我想要的，但是当我合并数百个表时需要很长时间。我觉得可能有一种更简单的方法可以做到，但我不想再花三天的时间来试错。

我想象的是这样的事情：

将空表 n 次导入数据帧列表
导入数据表并将其合并为列表中的空数据框
创建一个新数据框，该数据框是所有导入的数据框中相应单元格的平均值

有什么建议吗？

更新：以下是dput的示例表：

table.1 <- structure(list(A = c(12L, 62L, 37L, 74L), B = c(16L, 66L, 31L, 
76L), C = c(17L, 9L, 59L, 89L), D = c(14L, 85L, 75L, 25L)), .Names = c("A", 
"B", "C", "D"), row.names = c("A", "B", "C", "D"), class = "data.frame")

table.2 <- structure(list(A = c(12L, 62L, 24L, 7L), B = c(16L, 66L, 21L, 
1L), E = c(11L, 57L, 4L, 68L), F = c(19L, 28L, 51L, 22L)), .Names = c("A", 
"B", "E", "F"), row.names = c("A", "B", "E", "F"), class = "data.frame")

table.3 <- structure(list(C = c(59L, 89L, 67L, 39L), D = c(75L, 25L, 87L, 
69L), E = c(77L, 88L, 4L, 68L), F = c(80L, 30L, 51L, 24L)), .Names = c("C", 
"D", "E", "F"), row.names = c("C", "D", "E", "F"), class = "data.frame")

out.file <- structure(list(A = c(NA, NA, NA, NA, NA, NA), B = c(NA, NA, NA, 
NA, NA, NA), C = c(NA, NA, NA, NA, NA, NA), D = c(NA, NA, NA, 
NA, NA, NA), E = c(NA, NA, NA, NA, NA, NA), F = c(NA, NA, NA, 
NA, NA, NA)), .Names = c("A", "B", "C", "D", "E", "F"), row.names = c("A", 
"B", "C", "D", "E", "F"), class = "data.frame")

Answer 1

子集化解决方案，没有额外的包（使用@emehex定义的df1，df2和df3）：

# List of dataframes to combine
DF<-list(df1, df2, df3)

COL<-unique(unlist(lapply(DF, colnames)))
ROW<-unique(unlist(lapply(DF, rownames)))
# Empty DF with all combinations
TOTAL<-matrix(data=NA, nrow=length(ROW), ncol=length(COL), dimnames=list(ROW, COL))
# Subsetting :
for (df in DF) { 
    TOTAL[rownames(df), colnames(df)] <- as.matrix(df)
}

子集比合并更快，有许多数据帧可能更有效（参见@aichao回答她：For each row extract the value in the column name that match another value in the cell）。您只需根据代码调整DF列表file.names。

Answer 2

不知道你的.csvs是什么样的，所以这是我能做的最好的事情（上面有三个示例表）......

数据导入

"2001-07-30"

df1 <- read.table(header = TRUE, text = "A B C D A 12 16 17 14 B 62 66 9 85 C 37 31 59 75 D 74 76 89 25") df2 <- read.table(header = TRUE, text = "A B E F A 12 16 11 19 B 62 66 57 28 E 24 21 4 51 F 7 1 68 22") df3 <- read.table(header = TRUE, text = "C D E F C 59 75 77 80 D 89 25 88 30 E 67 87 4 51 F 39 69 68 22")，dplyr和tibble

的解决方案

tidyr

输出

library(dplyr)
library(tibble)
library(tidyr)

# intermediate tables for rownames and gathering
df1_c <- df1 %>% 
    rownames_to_column("Name") %>% 
    gather(key, value, -Name)

df2_c <- df2 %>% 
    rownames_to_column("Name") %>% 
    gather(key, value, -Name)

df3_c <- df3 %>% 
    rownames_to_column("Name") %>% 
    gather(key, value, -Name)

# formatted dataframe from spread
df <- bind_rows(df1_c, df2_c, df3_c) %>% 
    group_by(Name, key) %>% 
    distinct(.keep_all = TRUE) %>% 
    spread(key, value)

在R中按行和列合并多个表

2 个答案: