Question

我有多列的Excel表格。我需要创建具有最高列组值的新列。这需要针对每个列组（月）。我发布了这个问题，因为大多数先前的问题和答案都是关于单级列导入和重新整形。

Answer 1

如果可能，应避免在原始文件中合并单元格。如果您的xls文件不是太乱，那么可以处理：

library(readxl)
library(reshape)

# load an example excel file and see how it looks
d <- as.data.frame(read_excel("~/Bureau/Merged_Headers_Example.xls"))
d

##     X__1 January X__2 X__3 February X__4 X__5 March X__6 X__7
## 1 SECTOR    var1 var2 var3     var1 var2 var3  var1 var2 var3
## 2   AAAA       2    8    2        0    0    5     2    3    9
## 3   BBBB       9    3    8        4    2    6     8    6    8
## 4   CCCC       6    0   10        7    4    9     5    8    2
## 5   DDDD       0    1    2        2    5    7     4    3    2
## 6   EEEE       4    8    1        8   10    8     5   10    4
## 7   FFFF       1    0   10        2    1    4     1    3    3


# Capture the values of the merged cells on the first line
header <- colnames(d)
header <- header[!grepl("X__", header)]

# remove the first line and set the colomn names as the second line
name_cols <- d[1,]
d <- d[-1,]
colnames(d) <- name_cols

# Repeat the months the correct number of time 
# to be adapted depending on the structure of the file
# e.g. `-1` here is to not count the first column
header <- factor(rep(header, each = nrow(d) * (length(unique(colnames(d)))-1)),
                 levels = header)

# Dataset in "long" "tidy" form and add the monthes
d <- melt(d, id = "SECTOR")
d$month <- header

# Now you can rearange that as you want with cast
result <- cast(d, month + SECTOR ~ variable)
result

##       month SECTOR var1 var2 var3
## 1   January   AAAA    2    8    2
## 2   January   BBBB    9    3    8
## 3   January   CCCC    6    0   10
## 4   January   DDDD    0    1    2
## 5   January   EEEE    4    8    1
## 6   January   FFFF    1    0   10
## 7  February   AAAA    2    8    2
## 8  February   BBBB    9    3    8
## 9  February   CCCC    6    0   10
## 10 February   DDDD    0    1    2
## 11 February   EEEE    4    8    1
## 12 February   FFFF    1    0   10
## 13    March   AAAA    2    8    2
## 14    March   BBBB    9    3    8
## 15    March   CCCC    6    0   10
## 16    March   DDDD    0    1    2
## 17    March   EEEE    4    8    1
## 18    March   FFFF    1    0   10

# Or to be very close to what you asked (but this really not a good idea 
# in most circumstances I think)
result <- cast(d, SECTOR + month ~ variable)
result <- do.call(cbind.data.frame, split(result, result$month))
colnames(result) <- gsub("^.*\\.", "", colnames(result))
result

##    SECTOR   month var1 var2 var3 SECTOR    month var1 var2 var3 SECTOR
## 1    AAAA January    2    8    2   AAAA February    2    8    2   AAAA
## 4    BBBB January    9    3    8   BBBB February    9    3    8   BBBB
## 7    CCCC January    6    0   10   CCCC February    6    0   10   CCCC
## 10   DDDD January    0    1    2   DDDD February    0    1    2   DDDD
## 13   EEEE January    4    8    1   EEEE February    4    8    1   EEEE
## 16   FFFF January    1    0   10   FFFF February    1    0   10   FFFF
##    month var1 var2 var3
## 1  March    2    8    2
## 4  March    9    3    8
## 7  March    6    0   10
## 10 March    0    1    2
## 13 March    4    8    1
## 16 March    1    0   10

将多列Excel数据导入并重新整形为r

1 个答案: