自定义csv导入和合并功能，可用于r中的多个文件

Question

我有超过100个csv文件包含这样的数据...

> dput(head(hobo.temp))
structure(list(Serial = c("Plot Title: 20461693", "#", "1", 
"2", "3", "4"), Date = c("", "Date Time, GMT-05:00", "02/14/20 10:14:50 AM", 
"02/14/20 10:14:57 AM", "02/14/20 11:14:50 AM", "02/14/20 12:14:50 PM"
), Temp = c("", "Temp, °C (LGR S/N: 20461693, SEN S/N: 20461693)", 
"18.866", "", "20.817", "20.913"), X1 = c("", "Coupler Detached (LGR S/N: 20461693, SEN S/N: 20461693)", 
"", "Logged", "", ""), X2 = c("", "Coupler Attached (LGR S/N: 20461693, SEN S/N: 20461693)", 
"", "", "", ""), X3 = c("", "Host Connected (LGR S/N: 20461693, SEN S/N: 20461693)", 
"", "", "", ""), X4 = c("", "End Of File (LGR S/N: 20461693, SEN S/N: 20461693)", 
"", "", "", "")), row.names = c(NA, 6L), class = "data.frame")

它很脏，所以我写了代码来清理它。

hobo.temp <- read.csv("20461693_suw_main_01_19_2021.csv",
                      colClasses = c(rep("character", 3), rep("NULL", 4)),
                      col.names = c("Serial", "Date", "Temp", 1, 2, 3, 4), 
                      header = FALSE, fill = TRUE, stringsAsFactors = FALSE)
hobo.temp$Date = as.POSIXct(hobo.temp$Date,  format="%m/%d/%y %H:%M")
hobo.temp[,1] <- hobo.temp[1,1]
hobo.temp <- hobo.temp[-c(1:4),]
hobo.temp <- na.omit(hobo.temp)
hobo.temp <- arrange(hobo.temp, Date)
row.names(hobo.temp) <- NULL
hobo.temp$Serial <- gsub("Plot Title: ", "", hobo.temp$Serial, fixed = TRUE)
hobo.temp$Temp <- as.numeric(hobo.temp$Temp)
return(hobo.temp)

但是当我尝试将它转换为函数并使用下面的代码迭代它时。

filenames <- list.files(path = ".", pattern='^.*\\.csv$')

hobo.read <- function(fnam) {
  hobo.temp <- read.csv(fnam, colClasses = c(rep("character", 3), rep("NULL", 4)),
                        col.names = c("Serial", "Date", "Temp", 1, 2, 3, 4), 
                        header = FALSE, fill = TRUE, stringsAsFactors = FALSE)
  hobo.temp$Date = as.POSIXct(hobo.temp$Date,  format="%m/%d/%y %H:%M")
  hobo.temp[,1] <- hobo.temp[1,1]
  hobo.temp <- hobo.temp[-c(1:4),]
  hobo.temp <- na.omit(hobo.temp)
  hobo.temp <- arrange(hobo.temp, Date)
  row.names(hobo.temp) <- NULL
  hobo.temp$Serial <- gsub("Plot Title: ", "", hobo.temp$Serial, fixed = TRUE)
  hobo.temp$Temp <- as.numeric(hobo.temp$Temp)
  return(hobo.temp)
}

my.df <- do.call("rbind", lapply(filenames, hobo.read))

我得到了这个错误

 Error in read.table(file = file, header = header, sep = sep, quote = quote,  : 
  more columns than column names 

我不擅长写函数，所以我提前道歉。

Answer 1

我意识到有几个文件有8列。我认为这可能是这种情况，并试图用我原来的问题代码colClasses = c(rep("character", 3), rep("NULL", 4))来解释它。当我将`rep("NULL"，5)中的4切换为5时，它正确地使第8列无效。我修改了我原来的问题代码，使之更具可读性(也许是这样)。这是我的第一个真正的函数，它嵌套在boot中。这是草率的，但我对此感到非常自豪。

    #reads filenames from 
filenames <- list.files(path = ".", pattern='^.*\\.csv$')

#first function imports data
hobo.read <- function(x) {
  #in... rep("NULL", 5)... 5 has to be larger than the number for columns in the csv with the largest number of columns
  df1 <- read.csv(x, colClasses = c(rep("character", 3), rep("NULL", 5)),
                        col.names = c("Serial", "Date", "Temp", 1, 2, 3, 4), 
                        header = FALSE, fill = TRUE, stringsAsFactors = FALSE)
  # line applies the action function below
  df2 <- hobo.fix(df1)
  
}

#function of actions to apply within 1st function
hobo.fix <- function(hobo.temp) {
  hobo.temp[,1] <- hobo.temp[1,1]
  hobo.temp <- hobo.temp[-c(1:4),]
  hobo.temp$Serial <- gsub("Plot Title: ", "", hobo.temp$Serial, fixed = TRUE)
  hobo.temp$Temp <- as.numeric(hobo.temp$Temp)
  hobo.temp$Date = as.POSIXct(hobo.temp$Date,  format="%m/%d/%y %H:%M")
  hobo.temp <- na.omit(hobo.temp)
  hobo.temp <- dplyr::arrange(hobo.temp, Date)
  row.names(hobo.temp) <- NULL
  return(hobo.temp)
}


hobo <- do.call("rbind", lapply(filenames, hobo.read))