Learn R Programming

data.table (version 1.11.2)

fread: Fast and friendly file finagler

Description

Similar to read.table but faster and more convenient. All controls such as sep, colClasses and nrows are automatically detected. bit64::integer64 types are also detected and read directly without needing to read as character before converting.

Dates are read as character currently. They can be converted afterwards using the excellent fasttime package or standard base functions.

`fread` is for regular delimited files; i.e., where every row has the same number of columns. In future, secondary separator (sep2) may be specified within each column. Such columns will be read as type list where each cell is itself a vector.

Usage

fread(input, file, sep="auto", sep2="auto", dec=".", quote="\"",
nrows=Inf, header="auto",
na.strings=getOption("datatable.na.strings","NA"),  # due to change to ""; see NEWS
stringsAsFactors=FALSE, verbose=getOption("datatable.verbose", FALSE),
skip="__auto__", select=NULL, drop=NULL, colClasses=NULL,
integer64=getOption("datatable.integer64", "integer64"),
col.names,
check.names=FALSE, encoding="unknown",
strip.white=TRUE, fill=FALSE, blank.lines.skip=FALSE,
key=NULL, index=NULL,
showProgress=getOption("datatable.showProgress", interactive()),
data.table=getOption("datatable.fread.datatable", TRUE),
nThread=getDTthreads(),
logical01=getOption("datatable.logical01", FALSE),  # due to change to TRUE; see NEWS
autostart=NA
)

Arguments

input

Either the file name to read (containing no \n character), a shell command that pre-processes the file (e.g. fread("grep blah filename")) or the input itself as a string (containing at least one \n), see examples. In both cases, a length 1 character string. A filename input is passed through path.expand for convenience. input can also be a URL starting with http:// or file://; see Details.

sep

The separator between columns. Defaults to the character in the set [,\t |;:] that separates the sample of rows into the most number of lines with the same number of fields. Use NULL or "" to specify no separator; i.e. each line a single character column like base::readLines does.

sep2

The separator within columns. A list column will be returned where each cell is a vector of values. This is much faster using less working memory than strsplit afterwards or similar techniques. For each column sep2 can be different and is the first character in the same set above [,\t |;], other than sep, that exists inside each field outside quoted regions in the sample. NB: sep2 is not yet implemented.

nrows

The maximum number of rows to read. Unlike read.table, you do not need to set this to an estimate of the number of rows in the file for better speed because that is already automatically determined by fread almost instantly using the large sample of lines. `nrows=0` returns the column names and typed empty columns determined by the large sample; useful for a dry run of a large file or to quickly check format consistency of a set of files before starting to read any of them.

header

Does the first data line contain column names? Defaults according to whether every non-empty field on the first data line is type character. If so, or TRUE is supplied, any empty column names are given a default name.

na.strings

A character vector of strings which are to be interpreted as NA values. By default, ",," for columns of all types, including type `character` is read as NA for consistency. ,"", is unambiguous and read as an empty string. To read ,NA, as NA, set na.strings="NA". To read ,, as blank string "", set na.strings=NULL. When they occur in the file, the strings in na.strings should not appear quoted since that is how the string literal ,"NA", is distinguished from ,NA,, for example, when na.strings="NA".

file

File path, useful when we want to ensure that no shell commands will be executed. File path can also be provided to input argument.

stringsAsFactors

Convert all character columns to factors?

verbose

Be chatty and report timings?

skip

If 0 (default) start on the first line and from there finds the first row with a consistent number of columns. This automatically avoids irregular header information before the column names row. skip>0 means ignore the first skip rows manually. skip="string" searches for "string" in the file (e.g. a substring of the column names row) and starts on that line (inspired by read.xls in package gdata).

select

Vector of column names or numbers to keep, drop the rest.

drop

Vector of column names or numbers to drop, keep the rest.

colClasses

A character vector of classes (named or unnamed), as read.csv. Or a named list of vectors of column names or numbers, see examples. colClasses in fread is intended for rare overrides, not for routine use. fread will only promote a column to a higher type if colClasses requests it. It won't downgrade a column to a lower type since NAs would result. You have to coerce such columns afterwards yourself, if you really require data loss.

integer64

"integer64" (default) reads columns detected as containing integers larger than 2^31 as type bit64::integer64. Alternatively, "double"|"numeric" reads as base::read.csv does; i.e., possibly with loss of precision and if so silently. Or, "character".

dec

The decimal separator as in base::read.csv. If not "." (default) then usually ",". See details.

col.names

A vector of optional names for the variables (columns). The default is to use the header column if present or detected, or if not "V" followed by the column number. This is applied after check.names and before key and index.

check.names

default is FALSE. If TRUE then the names of the variables in the data.table are checked to ensure that they are syntactically valid variable names. If necessary they are adjusted (by make.names) so that they are, and also to ensure that there are no duplicates.

encoding

default is "unknown". Other possible options are "UTF-8" and "Latin-1". Note: it is not used to re-encode the input, rather enables handling of encoded strings in their native encoding.

quote

By default ("\""), if a field starts with a double quote, fread handles embedded quotes robustly as explained under Details. If it fails, then another attempt is made to read the field as is, i.e., as if quotes are disabled. By setting quote="", the field is always read as if quotes are disabled. It is not expected to ever need to pass anything other than \"\" to quote; i.e., to turn it off.

strip.white

default is TRUE. Strips leading and trailing whitespaces of unquoted fields. If FALSE, only header trailing spaces are removed.

fill

logical (default is FALSE). If TRUE then in case the rows have unequal length, blank fields are implicitly filled.

blank.lines.skip

logical, default is FALSE. If TRUE blank lines in the input are ignored.

key

Character vector of one or more column names which is passed to setkey. It may be a single comma separated string such as key="x,y,z", or a vector of names such as key=c("x","y","z"). Only valid when argument data.table=TRUE. Where applicable, this should refer to column names given in col.names.

index

Character vector or list of character vectors of one or more column names which is passed to setindexv. As with key, comma-separated notation like index="x,y,z" is accepted for convenience. Only valid when argument data.table=TRUE. Where applicable, this should refer to column names given in col.names.

showProgress

TRUE displays progress on the console if the ETA is greater than 3 seconds. It is produced in fread's C code where the very nice (but R level) txtProgressBar and tkProgressBar are not easily available.

data.table

TRUE returns a data.table. FALSE returns a data.frame.

nThread

The number of threads to use. Experiment to see what works best for your data on your hardware.

logical01

If TRUE a column containing only 0s and 1s will be read as logical, otherwise as integer.

autostart

Deprecated and ignored with warning. Please use skip instead.

Value

A data.table by default. A data.frame when argument data.table=FALSE; e.g. options(datatable.fread.datatable=FALSE).

Details

A sample of 10,000 rows is used for a very good estimate of column types. 100 contiguous rows are read from 100 equally spaced points throughout the file including the beginning, middle and the very end. This results in a better guess when a column changes type later in the file (e.g. blank at the beginning/only populated near the end, or 001 at the start but 0A0 later on). This very good type guess enables a single allocation of the correct type up front once for speed, memory efficiency and convenience of avoiding the need to set colClasses after an error. Even though the sample is large and jumping over the file, it is almost instant regardless of the size of the file because a lazy on-demand memory map is used. If a jump lands inside a quoted field containing newlines, each newline is tested until 5 lines are found following it with the expected number of fields. The lowest type for each column is chosen from the ordered list: logical, integer, integer64, double, character. Rarely, the file may contain data of a higher type in rows outside the sample (referred to as an out-of-sample type exception). In this event fread will automatically reread just those columns from the beginning so that you don't have the inconvenience of having to set colClasses yourself; particularly helpful if you have a lot of columns. Such columns must be read from the beginning to correctly distinguish "00" from "000" when those have both been interpreted as integer 0 due to the sample but 00A occurs out of sample. Set verbose=TRUE to see a detailed report of the logic deployed to read your file.

There is no line length limit, not even a very large one. Since we are encouraging list columns (i.e. sep2) this has the potential to encourage longer line lengths. So the approach of scanning each line into a buffer first and then rescanning that buffer is not used. There are no buffers used in fread's C code at all. The field width limit is limited by R itself: the maximum width of a character string (currently 2^31-1 bytes, 2GB).

The filename extension (such as .csv) is irrelevant for "auto" sep and sep2. Separator detection is entirely driven by the file contents. This can be useful when loading a set of different files which may not be named consistently, or may not have the extension .csv despite being csv. Some datasets have been collected over many years, one file per day for example. Sometimes the file name format has changed at some point in the past or even the format of the file itself. So the idea is that you can loop fread through a set of files and as long as each file is regular and delimited, fread can read them all. Whether they all stack is another matter but at least each one is read quickly without you needing to vary colClasses in read.table or read.csv.

If an empty line is encountered then reading stops there with warning if any text exists after the empty line such as a footer. The first line of any text discarded is included in the warning message. Unless, it is single-column input. In that case blank lines are significant (even at the very end) and represent NA in the single column. So that fread(fwrite(DT))==DT. This default behaviour can be controlled using blank.lines.skip=TRUE|FALSE.

Line endings: All known line endings are detected automatically: \n (*NIX including Mac), \r\n (Windows CRLF), \r (old Mac) and \n\r (just in case). There is no need to convert input files first. fread running on any architecture will read a file from any architecture. Both \r and \n may be embedded in character strings (including column names) provided the field is quoted.

Decimal separator and locale: fread(...,dec=",") should just work. fread uses C function strtod to read numeric data; e.g., 1.23 or 1,23. strtod retrieves the decimal separator (. or , usually) from the locale of the R session rather than as an argument passed to the strtod function. So for fread(...,dec=",") to work, fread changes this (and only this) R session's locale temporarily to a locale which provides the desired decimal separator.

On Windows, "French_France.1252" is tried which should be available as standard (any locale with comma decimal separator would suffice) and on unix "fr_FR.utf8" (you may need to install this locale on unix). fread() is very careful to set the locale back again afterwards, even if the function fails with an error. The choice of locale is determined by options()$datatable.fread.dec.locale. This may be a vector of locale names and if so they will be tried in turn until the desired dec is obtained; thus allowing more than two different decimal separators to be selected. This is a new feature in v1.9.6 and is experimental. In case of problems, turn it off with options(datatable.fread.dec.experiment=FALSE).

Quotes:

When quote is a single character,

  • Spaces and other whitespace (other than sep and \n) may appear in unquoted character fields, e.g., ...,2,Joe Bloggs,3.14,....

  • When character columns are quoted, they must start and end with that quoting character immediately followed by sep or \n, e.g., ...,2,"Joe Bloggs",3.14,....

    In essence quoting character fields are required only if sep or \n appears in the string value. Quoting may be used to signify that numeric data should be read as text. Unescaped quotes may be present in a quoted field, e.g., ...,2,"Joe, "Bloggs"",3.14,..., as well as escaped quotes, e.g., ...,2,"Joe \",Bloggs\"",3.14,....

    If an embedded quote is followed by the separator inside a quoted field, the embedded quotes up to that point in that field must be balanced; e.g. ...,2,"www.blah?x="one",y="two"",3.14,....

    On those fields that do not satisfy these conditions, e.g., fields with unbalanced quotes, fread re-attempts that field as if it isn't quoted. This is quite useful in reading files that contains fields with unbalanced quotes as well, automatically.

To read fields as is instead, use quote = "".

File Download:

When input begins with http://, https://, ftp://, ftps://, or file://, fread detects this and downloads the target to a temporary file (at tempfile()) before proceeding to read the file as usual. Secure URLS (ftps:// and https://) are downloaded with curl::curl_download; ftp:// and http:// paths are downloaded with download.file and method set to getOption("download.file.method"), defaulting to "auto"; and file:// is downloaded with download.file with method="internal". NB: this implies that for file://, even files found on the current machine will be "downloaded" (i.e., hard-copied) to a temporary file. See download.file for more details.

References

Background : https://cran.r-project.org/doc/manuals/R-data.html http://stackoverflow.com/questions/1727772/quickly-reading-very-large-tables-as-dataframes-in-r http://www.biostat.jhsph.edu/~rpeng/docs/R-large-tables.html http://www.cerebralmastication.com/2009/11/loading-big-data-into-r/ http://stackoverflow.com/questions/9061736/faster-than-scan-with-rcpp http://stackoverflow.com/questions/415515/how-can-i-read-and-manipulate-csv-file-data-in-c http://stackoverflow.com/questions/9352887/strategies-for-reading-in-csv-files-in-pieces http://stackoverflow.com/questions/11782084/reading-in-large-text-files-in-r http://stackoverflow.com/questions/45972/mmap-vs-reading-blocks http://stackoverflow.com/questions/258091/when-should-i-use-mmap-for-file-access http://stackoverflow.com/a/9818473/403310 http://stackoverflow.com/questions/9608950/reading-huge-files-using-memory-mapped-files

finagler = "to get or achieve by guile or manipulation" http://dictionary.reference.com/browse/finagler

See Also

read.csv, url, Sys.setlocale

Examples

Run this code
# NOT RUN {
# Demo speed-up
n = 1e6
DT = data.table( a=sample(1:1000,n,replace=TRUE),
                 b=sample(1:1000,n,replace=TRUE),
                 c=rnorm(n),
                 d=sample(c("foo","bar","baz","qux","quux"),n,replace=TRUE),
                 e=rnorm(n),
                 f=sample(1:1000,n,replace=TRUE) )
DT[2,b:=NA_integer_]
DT[4,c:=NA_real_]
DT[3,d:=NA_character_]
DT[5,d:=""]
DT[2,e:=+Inf]
DT[3,e:=-Inf]

write.table(DT,"test.csv",sep=",",row.names=FALSE,quote=FALSE)
cat("File size (MB):", round(file.info("test.csv")$size/1024^2),"\n")
# 50 MB (1e6 rows x 6 columns)

system.time(DF1 <-read.csv("test.csv",stringsAsFactors=FALSE))
# 60 sec (first time in fresh R session)

system.time(DF1 <- read.csv("test.csv",stringsAsFactors=FALSE))
# 30 sec (immediate repeat is faster, varies)

system.time(DF2 <- read.table("test.csv",header=TRUE,sep=",",quote="",
    stringsAsFactors=FALSE,comment.char="",nrows=n,
    colClasses=c("integer","integer","numeric",
                 "character","numeric","integer")))
# 10 sec (consistently). All known tricks and known nrows, see references.

require(data.table)
if(all(sapply(c("sqldf", "ff"), requireNamespace, quietly = TRUE))) {
  require(sqldf)
  require(ff)

  system.time(DT <- fread("test.csv"))
  #  3 sec (faster and friendlier)

  system.time(SQLDF <- read.csv.sql("test.csv",dbname=NULL))
  # 20 sec (friendly too, good defaults)

  system.time(FFDF <- read.csv.ffdf(file="test.csv",nrows=n))
  # 20 sec (friendly too, good defaults)

  identical(DF1,DF2)
  all.equal(as.data.table(DF1), DT)
  identical(DF1,within(SQLDF,{b<-as.integer(b);c<-as.numeric(c)}))
  identical(DF1,within(as.data.frame(FFDF),d<-as.character(d)))
}

# Scaling up ...
l = vector("list",10)
for (i in 1:10) l[[i]] = DT
DTbig = rbindlist(l)
tables()
write.table(DTbig,"testbig.csv",sep=",",row.names=FALSE,quote=FALSE)
# 500MB (10 million rows x 6 columns)

system.time(DF <- read.table("testbig.csv",header=TRUE,sep=",",
    quote="",stringsAsFactors=FALSE,comment.char="",nrows=1e7,
    colClasses=c("integer","integer","numeric",
                 "character","numeric","integer")))
# 100-200 sec (varies)

system.time(DT <- fread("testbig.csv"))
# 30-40 sec

all(mapply(all.equal, DF, DT))


# Real data example (Airline data)
# http://stat-computing.org/dataexpo/2009/the-data.html

download.file("http://stat-computing.org/dataexpo/2009/2008.csv.bz2",
              destfile="2008.csv.bz2")
# 109MB (compressed)

system("bunzip2 2008.csv.bz2")
# 658MB (7,009,728 rows x 29 columns)

colClasses = sapply(read.csv("2008.csv",nrows=100),class)
# 4 character, 24 integer, 1 logical. Incorrect.

colClasses = sapply(read.csv("2008.csv",nrows=200),class)
# 5 character, 24 integer. Correct. Might have missed data only using 100 rows
# since read.table assumes colClasses is correct.

system.time(DF <- read.table("2008.csv", header=TRUE, sep=",",
    quote="",stringsAsFactors=FALSE,comment.char="",nrows=7009730,
    colClasses=colClasses))
# 360 secs

system.time(DT <- fread("2008.csv"))
#  40 secs

table(sapply(DT,class))
# 5 character and 24 integer columns. Correct without needing to worry about colClasses
# issue above.


# Reads URLs directly :
fread("http://www.stats.ox.ac.uk/pub/datasets/csb/ch11b.dat")

# }
# NOT RUN {
# Reads text input directly :
fread("A,B\n1,2\n3,4")

# Reads pasted input directly :
fread("A,B
1,2
3,4
")

# Finds the first data line automatically :
fread("
This is perhaps a banner line or two or ten.
A,B
1,2
3,4
")

# Detects whether column names are present automatically :
fread("
1,2
3,4
")

# Numerical precision :

DT = fread("A\n1.010203040506070809010203040506\n")
# TODO: add numerals=c("allow.loss", "warn.loss", "no.loss") from base::read.table, +"use.Rmpfr"
typeof(DT$A)=="double"   # currently "allow.loss" with no option

DT = fread("A\n1.46761e-313\n")   # read as 'numeric'
DT[,sprintf("%.15E",A)]   # beyond what double precision can store accurately to 15 digits
# For greater accuracy use colClasses to read as character, then package Rmpfr.

# colClasses
data = "A,B,C,D\n1,3,5,7\n2,4,6,8\n"
fread(data, colClasses=c(B="character",C="character",D="character"))  # as read.csv
fread(data, colClasses=list(character=c("B","C","D")))    # saves typing
fread(data, colClasses=list(character=2:4))     # same using column numbers

# drop
fread(data, colClasses=c("B"="NULL","C"="NULL"))   # as read.csv
fread(data, colClasses=list(NULL=c("B","C")))      #
fread(data, drop=c("B","C"))      # same but less typing, easier to read
fread(data, drop=2:3)             # same using column numbers

# select
# (in read.csv you need to work out which to drop)
fread(data, select=c("A","D"))    # less typing, easier to read
fread(data, select=c(1,4))        # same using column numbers

# skip blank lines
fread("a,b\n1,a\n2,b\n\n\n3,c\n", blank.lines.skip=TRUE)
# fill
fread("a,b\n1,a\n2\n3,c\n", fill=TRUE)
fread("a,b\n\n1,a\n2\n\n3,c\n\n", fill=TRUE)

# fill with skip blank lines
fread("a,b\n\n1,a\n2\n\n3,c\n\n", fill=TRUE, blank.lines.skip=TRUE)

# check.names usage
fread("a b,a b\n1,2\n")
fread("a b,a b\n1,2\n", check.names=TRUE) # no duplicates + syntactically valid names
# }

Run the code above in your browser using DataLab