read.csv warning 'EOF within quoted string' prevents complete reading of file
I have a CSV file (24.1 MB) that I cannot fully read into my R session. When I open the file in a spreadsheet program I can see 112,544 rows. When I read it into R with read.csv I only get 56,952 rows and this warning:
cit <- read.csv("citations.CSV", row.names = NULL,
comment.char = "", header = TRUE,
stringsAsFactors = FALSE,
colClasses= "character", encoding= "utf-8")
Warning message:
In scan(file, what, nmax, sep, dec, quote, skip, nlines, na.strings, :
EOF within quoted string
I can read the whole file into R with readLines:
rl <- readLines(file("citations.CSV", encoding = "utf-8"))
length(rl)
[1] 112545
But I can't get this back into R as a table (via read.csv):
write.table(rl, "rl.txt", quote = FALSE, row.names = FALSE)
rl_in <- read.csv("rl.txt", skip = 1, row.names = NULL)
Warning message:
In scan(file, what, nmax, sep, dec, quote, skip, nlines, na.strings, :
EOF within quoted string
How can I solve or workaround this EOF message (which seems to be more of an error than a warning) to get the entire file into my R session?
I have similar problems with other methods of reading CSV files:
require(sqldf)
cit_sql <- read.csv.sql("citations.CSV", sql = "select * from file")
require(data.table)
cit_dt <- fread("citations.CSV")
require(ff)
cit_ff <- read.csv.ffdf(file="citations.CSV")
Here's my sessionInfo()
R version 3.0.1 (2013-05-16)
Platform: x86_64-w64-mingw32/x64 (64-bit)
locale:
[1] LC_COLLATE=English_United States.1252 LC_CTYPE=English_United States.1252
[3] LC_MONETARY=English_United States.1252 LC_NUMERIC=C
[5] LC_TIME=English_United States.1252
attached base packages:
[1] tools tcltk stats graphics grDevices utils datasets methods base
other attached packages:
[1] ff_2.2-11 bit_1.1-10 data.table_1.8.8 sqldf_0.4-6.4
[5] RSQLite.extfuns_0.0.1 RSQLite_0.11.4 chron_2.3-43 gsubfn_0.6-5
[9] proto_0.3-10 DBI_0.2-7
Answer
You need to disable quoting.
cit <- read.csv("citations.CSV", quote = "",
row.names = NULL,
stringsAsFactors = FALSE)
str(cit)
## 'data.frame': 112543 obs. of 13 variables:
## $ row.names : chr "10.2307/675394" "10.2307/30007362" "10.2307/4254931" "10.2307/20537934" ...
## $ id : chr "10.2307/675394\t" "10.2307/30007362\t" "10.2307/4254931\t" "10.2307/20537934\t" ...
## $ doi : chr "Archaeological Inference and Inductive Confirmation\t" "Sound and Sense in Cath Almaine\t" "Oak Galls Preserved by the Eruption of Mount Vesuvius in A.D. 79_ and Their Probable Use\t" "The Arts Four Thousand Years Ago\t" ...
## $ title : chr "Bruce D. Smith\t" "Tomás Ó Cathasaigh\t" "Hiram G. Larew\t" "\t" ...
## $ author : chr "American Anthropologist\t" "Ériu\t" "Economic Botany\t" "The Illustrated Magazine of Art\t" ...
## $ journaltitle : chr "79\t" "54\t" "41\t" "1\t" ...
## $ volume : chr "3\t" "\t" "1\t" "3\t" ...
## $ issue : chr "1977-09-01T00:00:00Z\t" "2004-01-01T00:00:00Z\t" "1987-01-01T00:00:00Z\t" "1853-01-01T00:00:00Z\t" ...
## $ pubdate : chr "pp. 598-617\t" "pp. 41-47\t" "pp. 33-40\t" "pp. 171-172\t" ...
## $ pagerange : chr "American Anthropological Association\tWiley\t" "Royal Irish Academy\t" "New York Botanical Garden Press\tSpringer\t" "\t" ...
## $ publisher : chr "fla\t" "fla\t" "fla\t" "fla\t" ...
## $ type : logi NA NA NA NA NA NA ...
## $ reviewed.work: logi NA NA NA NA NA NA ...
I think is because of this kind of lines (check "Thorn" and "Minus")
readLines("citations.CSV")[82]
[1] "10.2307/3642839,10.2307/3642839\t,\"Thorn\" and \"Minus\" in Hieroglyphic Luvian Orthography\t,H. Craig Melchert\t,Anatolian Studies\t,38\t,\t,1988-01-01T00:00:00Z\t,pp. 29-42\t,British Institute at Ankara\t,fla\t,\t,"