Fast data reading with fread()
Data Manipulation with data.table in R
Matt Dowle, Arun Srinivasan
Instructors, DataCamp
Blazing FAST!
- Fast and parallel file reader
- Argument
nThreadcontrols the number of threads to use
User-friendly
- Can import local files, files from the web, and strings
- Intelligent defaults -
colClasses,sep,nrowsetc. - Note: Dates and Datetimes are read as character columns but can be converted later with the excellent
fasttimeoranytimepackages
Fast and friendly file reader
# File from URL
DT1<-fread("https://bit.ly/2RkBXhV")
DT1
a b
1 2
3 4
# Local file
DT2 <- fread("data.csv")
DT2
a b
1 2
3 4
# String
DT3 <- fread("a,b\n1,2\n3,4")
DT3
a b
1 2
3 4
# String without col names
DT4 <- fread("1,2\n3,4")
DT4
V1 V2
1 2
3 4
nrows and skip arguments
# Read only first line (after header)
fread("a,b\n1,2\n3,4", nrows = 1)
a b
1 2
# Skip first two lines containing metadata
str <- "# Metadata\nTimestamp: 2018-05-01 19:44:28 GMT\na,b\n1,2\n3,4"
fread(str, skip = 2)
a b
1 2
3 4
More on nrows and skip arguments
str <- "# Metadata\nTimestamp: 2018-05-01 19:44:28 GMT\na,b\n1,2\n3,4"
fread(str, skip = "a,b")
a b
1 2
3 4
fread(str, skip = "a,b", nrows = 1)
a b
1 2
select and drop arguments
str <- "a,b,c\n1,2,x\n3,4,y"
fread(str, select = c("a", "c"))
# Same as
fread(str, drop = "b")
a c
1 x
3 y
str <- "1,2,x\n3,4,y"
fread(str, select = c(1, 3))
# Same as
fread(str, drop = 2)
V1 V3
1 x
3 y
Let's practice!
Data Manipulation with data.table in R
