读取带有名称和标签的.csv文件到R中

Question

I have a .csv file that I need to read into R. The first row contains the names (eg BFI1, BFI2, CAQ2) and the the second row contains the question which I would also like to access in R (eg "I enjoy going to parties"). Each row after the first two corresponds to one participant.

I would like to be able to access the Codes and the text in R. (eg to use grep to access all the questions from one survey, and also to see the item text if needed. I need the numerical responses to be numeric.

BFI1, BFI2, CAQ1, CAQ2
Likes to read, Enjoys Parties, Is Nervous, Loves Books
3, 7, 1, 4
4, 5, 3, 3

I want to read this in so that I can access either the names (row 1) or the text (as labels maybe). I have looked at the Hmisc package but their label functionality seems limited.

Is there any way to read in this .csv file and access both of these values?

Answer 1

Not sure if you're okay with having the labels as a separate vector, but here's an idea. Suppose your file name is x.txt

## set up an argument list for scan() - just to avoid repetition
scanArgs <- list(
    file = "x.txt", what = "", nlines = 1, sep = ",", strip.white = TRUE
)

## read the data with no header and add the first line as names
df <- setNames(
    read.table("x.txt", skip = 2, sep = ","), 
    do.call(scan, scanArgs)
)
#   BFI1 BFI2 CAQ1 CAQ2
# 1    3    7    1    4
# 2    4    5    3    3

## make the label vector
labels <- setNames(do.call(scan, c(scanArgs, skip = 1)), names(df))
#            BFI1             BFI2             CAQ1             CAQ2 
# "Likes to read" "Enjoys Parties"     "Is Nervous"    "Loves Books" 

So the elements in labels correspond to the columns in df and the columns are numeric.

Note that x.txt was created with

txt <- 'BFI1, BFI2, CAQ1, CAQ2
Likes to read, Enjoys Parties, Is Nervous, Loves Books
3,7,1,4
4,5,3,3'
writeLines(txt, "x.txt")

Answer 2

You can use the nrows and skip arguments or read.csv

nameFile <- "data.csv"

# read the first two lines
vectorNames <- read.csv(nameFile, nrows = 1)
vectorDescription <- read.csv(nameFile, nrows = 1, skip = 1)

# read the data
dfIn <- read.csv(nameFile, skip = 2)
names(dfIn) <- vectorNames

Answer 3

@Richard Scriven I used your code and followed it up with this using the package

library(Hmisc)
y=data.frame(temp=rep(NA,nrow(df)))  
for (i in 1:length(labels)){  
x=df[,i]  
label(x)=labels[i]   
y[names(df)[i]]=x  
}  
y$temp=NULL  
y  
#  BFI1 BFI2 CAQ1 CAQ2
# 1    3    7    1    4
# 2    4    5    3    3
label(y)
#            BFI1             BFI2             CAQ1             CAQ2 
# "Likes to read" "Enjoys Parties"     "Is Nervous"    "Loves Books" 

Answer 4

Building on Michelle Usuelli's answer and with Rich Scriven correction, you can write this function:

read_csv_with_labels <- function(fileName)
{
 library(Hmisc)

 # read the first two lines
 varNames <- read.csv(fileName, nrows = 1, stringsAsFactors = FALSE, header = FALSE)
 varLabels <- read.csv(fileName, nrows = 1, stringsAsFactors = FALSE, header = TRUE)

 # read the data
 df <- read.csv(fileName, skip = 2)

 # assign variable names and labels to the dataframe
 names(df) <- varNames
 label(df) <- varLabels 

 return(df)
}

I think this should be included in the basic functionality of read.csv and read_csv.