How to read multiple xlsx file in R using loop wit

问题:

I have to read multiple xlsx file with random names into single dataframe. Structure of each file is same. I have to import specific columns only.

I tried this:

dat <- read.xlsx("FILE.xlsx", sheetIndex=1, 
                  sheetName=NULL, startRow=5, 
                  endRow=NULL, as.data.frame=TRUE, 
                  header=TRUE)

But this is for only one file at a time and I couldn't specify my particular columns. I even tried :

site=list.files(pattern='[.]xls')

but after that loop isn't working. How to do it? Thanks in advance.

回答1:

I would read each sheet to a list:

Get file names:

f = list.files("./")

Read files:

dat = lapply(f, function(i){
    x = read.xlsx(i, sheetIndex=1, sheetName=NULL, startRow=5,
        endRow=NULL, as.data.frame=TRUE, header=T)
    # Get the columns you want, e.g. 1, 3, 5
    x = x[, c(1, 3, 5)]
    # You may want to add a column to say which file they're from
    x$file = i
    # Return your data
    x
})

You can then access the items in your list with:

dat[[1]]

Or do the same task to them with:

lapply(dat, colmeans)

Turn them into a data frame (where your file column now becomes useful):

dat = do.call("rbind.data.frame", dat)

回答2:

I am more familiar with a for loop, which can be a bit more cumbersome.

filelist <- list.files(pattern = "\\.xlsx") # list all the xlsx files from the directory

allxlsx.files <- list()  # create a list to populate with xlsx data (if you wind to bind all the rows together)
count <- 1
for (file in filelist) {
   dat <- read.xlsx(file, sheetIndex=1, 
              sheetName=NULL, startRow=5, 
              endRow=NULL, as.data.frame=TRUE, 
              header=TRUE) [c(5:10, 12,15)] # index your columns of interest
   allxlsx.files[[count]] <-dat # creat a list of rows from xls files
   count <- count + 1
}

convert back to data.frame

allfiles <- do.call(rbind.data.frame, allxlsx.files)

回答3:

For a variation on Wyldsoul's answer, but using a for loop across multiple Excel sheets (between 1 and j) in the same Excel file, and binding with dplyr:

library(gdata) 
library(dplyr)

for (i in 1:j) {
  dat <- read.xls(f, sheet = i) 
  dat <- dat[,1:14] # index your columns of interest
  allxlsx.files[[count]]
  count <- count + 1
}

allfiles <- do.call(bind_rows, allxlsx.files)