Data loss during read.csv in R - r

Question

I have a .csv file to be imported into R, which has more than 1K observations. However, when I used the read.csv function as usual, the imported file only has 21 observations. This is strange. I've never seen this before.
t <- read.csv("E:\\AH1_09182014.CSV",header=T, colClasses=c(rep("character",3),rep("numeric",22)),na.string=c("null","NaN",""),stringsAsFactors=FALSE)
Can anyone help me figure out the problem? I am giving a link to my data file:
https://drive.google.com/file/d/0B86_a8ltyoL3TzBza0x1VTd2OTQ/edit?usp=sharing

Answer 1

You have some messy characters in your data--things like embedded control characters.
A workaround is to read the file in binary mode, and use read.csv on the text file read in.
This answer proposes a basic function to do those steps.
The function looks like this:
sReadLines <- function(fnam) {
f <- file(fnam, "rb")
res <- readLines(f)
close(f)
res
}
You can use it as follows:
temp <- read.csv(text = sReadLines("~/Downloads/AH1_09182014.CSV"),
stringsAsFactors = FALSE)
Have all lines been read in?
dim(temp)
# [1] 1449 25
Where is that problem line?
unlist(temp[21, ], use.names = FALSE)
# [1] "A-H Log 1" "09/18/2014" "0:19:00" "7.866" "255" "0.009"
# [7] "525" "7" "4468" "76" "4576.76" "20"
# [13] "71" "19" "77" "1222" "33857" "-3382"
# [19] "26\032)" "18.30" "84.80" "991.43" "23713.90" "0.85"
# [25] "10.54"
^^ see item [19] above.
Because of this, you won't be able to specify all of your column types up front--unless you clean the CSV first.