相比dplyr包，data.table包能够更大程度地提高数据的处理速度，这里就简单介绍一下data.tale包的使用方法。
数据的读取
data.table包中数据读取的函数是fread()。
data.table 的创建
library(data.table)
DT = data.table(x=rep(c("a","b","c"),each=3), y=c(1,3,6), v=1:9)
DT
# x y v
# 1: a 1 1
# 2: a 3 2
# 3: a 6 3
# 4: b 1 4
# 5: b 3 5
# 6: b 6 6
# 7: c 1 7
# 8: c 3 8
# 9: c 6 9
基础操作
行提取
行提取分为单行提取和多行提取。
单行提取
DT[2] # 2nd row
# x y v
# 1: a 3 2
DT[2,] # same
# x y v
# 1: a 3 2
这里DT[2]和DT[2,]是完全相同的，这里的『,』只是说明还有其他参数可设置，而其他参数按默认值进行计算。下文所有这样的最后一个『,』都不再写出来。
多行提取
- 数字提取
DT[1:2]
# x y v
# 1: a 1 1
# 2: a 3 2
DT[c(2,5)]
# x y v
#1: a 3 2
#2: b 3 5
- 逻辑提取
DT[c(FALSE,TRUE)] # even rows (usual recycling)
# x y v
# 1: a 3 2
# 2: b 1 4
# 3: b 6 6
# 4: c 3 8
此时,c(FALSE,TRUE)会自己重复匹配成与DT的行数相同的向量
列提取
与行提取相同，列的提取也包含单列提取和多列提取。
单列提取
- 数字提取
数字提取时，一定要把wth参数设置为FALSE。
DT[,2,with=FALSE] # 2nd column
# y
# 1: 1
# 2: 3
# 3: 6
# 4: 1
# 5: 3
# 6: 6
# 7: 1
# 8: 3
# 9: 6
- 列名提取
DT[,list(v)] # v column (as data.table
# v
# 1: 1
# 2: 2
# 3: 3
# 4: 4
# 5: 5
# 6: 6
# 7: 7
# 8: 8
# 9: 9
列名的修改
列名的修改可以使用setnames()函数，这个函数好像比对data.frame类型数据名更改的names()和colnames()函数也要快一些。
dt = data.table(a=1:2,b=3:4,c=5:6) # compare to data.table
try(tracemem(dt)) # by reference, no deep or shallow copies
setnames(dt,"b","B") # by name, no match() needed (warning if "b" is missing)
setnames(dt,3,"C") # by position with warning if 3 > ncol(dt)
setnames(dt,2:3,c("D","E")) # multiple
setnames(dt,c("a","E"),c("A","F")) # multiple by name (warning if either "a" or "E" is missing)
setnames(dt,c("X","Y","Z")) # replace all (length of names must be == ncol(DT))
多列提取
- 数字提取
如同上面对按数字对单列的提取，对多列提取也要设置with参数为FALSE。
DT[,2:3,with=FALSE]
# y v
# 1: 1 1
# 2: 3 2
# 3: 6 3
# 4: 1 4
# 5: 3 5
# 6: 6 6
# 7: 1 7
# 8: 3 8
# 9: 6 9
DT[,c(1,3),with=FALSE]
# x v
# 1: a 1
# 2: a 2
# 3: a 3
# 4: b 4
# 5: b 5
# 6: b 6
# 7: c 7
# 8: c 8
# 9: c 9
- 按列名提取
DT[,list(y, v)]
# y v
# 1: 1 1
# 2: 3 2
# 3: 6 3
# 4: 1 4
# 5: 3 5
# 6: 6 6
# 7: 1 7
# 8: 3 8
# 9: 6 9
如果按列名提取时，不使用列表，仍然能对列进行提取，只是结果以向量的形式输出。
DT[,v] # v column (as vector)
# [1] 1 2 3 4 5 6 7 8 9
DT[,c(v)] # same
# [1] 1 2 3 4 5 6 7 8 9
DT[, c(y, v)]
# [1] 1 3 6 1 3 6 1 3 6 1 2 3 4 5 6 7 8 9
列的添加与删除
列的添加
- 单列添加
DT
# x y v
# 1: a 1 1
# 2: a 3 2
# 3: a 6 3
# 4: b 1 4
# 5: b 3 5
# 6: b 6 6
# 7: c 1 7
# 8: c 3 8
# 9: c 6 9
DT[, a := 'k']
DT
# x y v a
# 1: a 1 1 k
# 2: a 3 2 k
# 3: a 6 3 k
# 4: b 1 4 k
# 5: b 3 5 k
# 6: b 6 6 k
# 7: c 1 7 k
# 8: c 3 8 k
# 9: c 6 9 k
DT[,c:=8] # add a numeric column, 8 for all rows
DT
# x y v a c
# 1: a 1 1 k 8
# 2: a 3 2 k 8
# 3: a 6 3 k 8
# 4: b 1 4 k 8
# 5: b 3 5 k 8
# 6: b 6 6 k 8
# 7: c 1 7 k 8
# 8: c 3 8 k 8
# 9: c 6 9 k 8
DT[,d:=9L] # add an integer column, 9L for all rows
DT[2,d:=10L] # subassign by reference to column d
DT
# x y v a c d
# 1: a 1 1 k 8 9
# 2: a 3 2 k 8 10
# 3: a 6 3 k 8 9
# 4: b 1 4 k 8 9
# 5: b 3 5 k 8 9
# 6: b 6 6 k 8 9
# 7: c 1 7 k 8 9
# 8: c 3 8 k 8 9
# 9: c 6 9 k 8 9
DT[, e := d + 2]
DT
# x y v a c d e
# 1: a 1 1 k 8 9 11
# 2: a 3 2 k 8 10 12
# 3: a 6 3 k 8 9 11
# 4: b 1 4 k 8 9 11
# 5: b 3 5 k 8 9 11
# 6: b 6 6 k 8 9 11
# 7: c 1 7 k 8 9 11
# 8: c 3 8 k 8 9 11
# 9: c 6 9 k 8 9 11
如果添加的列名，数据中已经包含则是对这一列数据的修改。
- 多列的添加
DT[, c('f', 'g') := list( d + 1, c)]
DT[, ':='( f = d + 1, g = c)] # same
DT
# x y v a c d e f g
# 1: a 1 1 k 8 9 11 10 8
# 2: a 3 2 k 8 10 12 11 8
# 3: a 6 3 k 8 9 11 10 8
# 4: b 1 4 k 8 9 11 10 8
# 5: b 3 5 k 8 9 11 10 8
# 6: b 6 6 k 8 9 11 10 8
# 7: c 1 7 k 8 9 11 10 8
# 8: c 3 8 k 8 9 11 10 8
# 9: c 6 9 k 8 9 11 10 8
此处，需要注意的是新创建的列只能依照原有数据列，而不能依照新创建的列。例如这个例子中，g = c是可以运行，而g = f则会提示错误。
列的删除
DT[,c:=NULL] # remove column c
DT
# x y v a d e f g
# 1: a 1 1 k 9 11 10 8
# 2: a 3 2 k 10 12 11 8
# 3: a 6 3 k 9 11 10 8
# 4: b 1 4 k 9 11 10 8
# 5: b 3 5 k 9 11 10 8
# 6: b 6 6 k 9 11 10 8
# 7: c 1 7 k 9 11 10 8
# 8: c 3 8 k 9 11 10 8
# 9: c 6 9 k 9 11 10 8
DT[, c('d', 'e', 'f', 'g'):=NULL]
DT
# x y v a
# 1: a 1 1 k
# 2: a 3 2 k
# 3: a 6 3 k
# 4: b 1 4 k
# 5: b 3 5 k
# 6: b 6 6 k
# 7: c 1 7 k
# 8: c 3 8 k
# 9: c 6 9 k
列指标的简单操作
简单操作主要包括求和、平均值、方差和标准差等。
DT[2:3,sum(v)] # sum(v) over rows 2 and 3
# [1] 5
DT[2:3,mean(v)] # sum(v) over rows 2 and 3
# [1] 2.5
索引key
查看和创建索引
索引是对列而言的，索引创建后，数据将自动按索引值进行重新排序，所以每个数据最多只能有一个索引，但是索引可以由多列组成，这些列可以是数字、因子、字符串或其他格式。
单列索引的创建
## methdod first
key(DT) # key
# NULL
setkey(DT,x) # set a 1-column key. No quotes, for convenience.
key(DT)
[1] "x"
DT
# x y v a
# 1: a 1 1 k
# 2: a 3 2 k
# 3: a 6 3 k
# 4: b 1 4 k
# 5: b 3 5 k
# 6: b 6 6 k
# 7: c 1 7 k
# 8: c 3 8 k
# 9: c 6 9 k
## method second
setkeyv(DT,"y") # same (v in setkeyv stands for vector)
key(DT)
# [1] "y"
一旦对数据进行新的索引，原有的索引将消失。
多列索引的创建
## methdod first # key
setkey(DT,x,v) # set a 1-column key. No quotes, for convenience.
key(DT)
# [1] "x" "v"
DT
# x y v a
# 1: a 1 1 k
# 2: a 3 2 k
# 3: a 6 3 k
# 4: b 1 4 k
# 5: b 3 5 k
# 6: b 6 6 k
# 7: c 1 7 k
# 8: c 3 8 k
# 9: c 6 9 k
## method second
setkeyv(DT,c("x", "y")) # same (v in setkeyv stands for vector)
key(DT)
# [1] "x" "v"
DT
# x y v a
# 1: a 1 1 k
# 2: a 3 2 k
# 3: a 6 3 k
# 4: b 1 4 k
# 5: b 3 5 k
# 6: b 6 6 k
# 7: c 1 7 k
# 8: c 3 8 k
# 9: c 6 9 k
通过索引进行数据的提取
按照索引对数据提取，可以加快提取数据的速度。
单索引
正向提取
setkey(DT, x)
DT["a"] # binary search (fast)
# x y v a
# 1: a 1 1 k
# 2: a 3 2 k
# 3: a 6 3 k
DT[.(x=="a")] # same; i.e. binary search (fast)
# x y v a
# 1: a 1 1 k
# 2: a 3 2 k
# 3: a 6 3 k
DT[x=="a"] # same; i.e. binary search (fast)
# x y v a
# 1: a 1 1 k
# 2: a 3 2 k
# 3: a 6 3 k
- 反向提取
DT[!.("a")] # not join
# x y v a
# 1: b 1 4 k
# 2: b 3 5 k
# 3: b 6 6 k
# 4: c 1 7 k
# 5: c 3 8 k
# 6: c 6 9 k
DT[!"a"] # same
# x y v a
# 1: b 1 4 k
# 2: b 3 5 k
# 3: b 6 6 k
# 4: c 1 7 k
# 5: c 3 8 k
# 6: c 6 9 k
DT[!2:4] # all rows other than 2:4
# x y v a
# 1: a 1 1 k
# 2: b 3 5 k
# 3: b 6 6 k
# 4: c 1 7 k
# 5: c 3 8 k
# 6: c 6 9 k
多索引
- 正向提取
setkey(DT, x, y)
# Mehtod First
DT["a"] # join to 1st column of key
# x y v a
# 1: a 1 1 k
# 2: a 3 2 k
# 3: a 6 3 k
DT[.("a")] # same, .() is an alias for list()
# x y v a
# 1: a 1 1 k
# 2: a 3 2 k
# 3: a 6 3 k
DT[.("a",3)] # join to 2 columns
# x y v a
# 1: a 3 2 k
DT[.("a",3:6)] # join 4 rows (2 missing)
# x y v a
# 1: a 3 2 k
# 2: a 4 NA NA
# 3: a 5 NA NA
# 4: a 6 3 k
DT[.("a",3:6),nomatch=0] # remove missing
# x y v a
# 1: a 3 2 k
# 2: a 6 3 k
DT[.("a",3:6),roll=TRUE] # rolling join (locf)
# x y v a
# 1: a 3 2 k
# 2: a 4 2 k
# 3: a 5 2 k
# 4: a 6 3 k
## Method Second
DT[J('a')]
# x y v a
# 1: a 1 1 k
# 2: a 3 2 k
# 3: a 6 3 k
DT[J("a",3)] # binary search (fast)
# x y v a
# 1: a 3 2 k
DT[J("a",3:6)] # same; i.e. binary search (fast)
# x y v a
# 1: a 3 2 k
# 2: a 4 NA NA
# 3: a 5 NA NA
# 4: a 6 3 k
DT[J("a",3:6), nomatch = 0]
# x y v a
# 1: a 3 2 k
# 2: a 6 3 k
DT[J("a",3:6), roll = T]
# x y v a
# 1: a 3 2 k
# 2: a 4 2 k
# 3: a 5 2 k
# 4: a 6 3 k
## Method Third
DT[list("a")]
# x y v a
# 1: a 1 1 k
# 2: a 3 2 k
# 3: a 6 3 k
DT[list("a",3)]
# x y v a
# 1: a 3 2 k
DT[list("a", 3:6)]
# x y v a
# 1: a 3 2 k
# 2: a 4 NA NA
# 3: a 5 NA NA
# 4: a 6 3 k
DT[list("a", 3:6), nomatch = 0]
# x y v a
# 1: a 3 2 k
# 2: a 6 3 k
DT[list("a", 3:6), roll = T]
# x y v a
# 1: a 3 2 k
# 2: a 4 2 k
# 3: a 5 2 k
# 4: a 6 3 k
- 反向提取
DT[x!="b" | y!=3] # not yet optimized, currently vector scans
# x y v a
# 1: a 1 1 k
# 2: a 3 2 k
# 3: a 6 3 k
# 4: b 1 4 k
# 5: b 6 6 k
# 6: c 1 7 k
# 7: c 3 8 k
# 8: c 6 9 k
DT[!.("b",3)] # same result but much faster
# x y v a
# 1: a 1 1 k
# 2: a 3 2 k
# 3: a 6 3 k
# 4: b 1 4 k
# 5: b 6 6 k
# 6: c 1 7 k
# 7: c 3 8 k
# 8: c 6 9 k
分类汇总
分类汇总是指按某列的分类指标进行简单操作，这里借助by参数实现。此外，by参数与索引相互没有影响。
单指标的分类汇总
- 默认汇总名称
DT[,sum(v),by=x]
# x V1
# 1: a 6
# 2: b 15
# 3: c 24
DT[,sum(v),by=y]
# y V1
# 1: 1 12
# 2: 3 15
# 3: 6 18
- 自定义汇总名称
DT[,list(sum.v.x = sum(v)),by=x]
# x sum.v.x
# 1: a 6
# 2: b 15
# 3: c 24
DT[,list(sum.v.y = sum(v)),by=y]
# y sum.v.y
# 1: 1 12
# 2: 3 15
# 3: 6 18
DT[,sum.v.y := sum(v) ,by=y]
# x y v a sum.v.y
# 1: a 1 1 k 12
# 2: a 3 2 k 15
# 3: a 6 3 k 18
# 4: b 1 4 k 12
# 5: b 3 5 k 15
# 6: b 6 6 k 18
# 7: c 1 7 k 12
# 8: c 3 8 k 15
# 9: c 6 9 k 18
- 汇总结果与原始数据进行匹配
DT[,sum.v.y := sum(v) ,by=y]
# x y v a sum.v.y
# 1: a 1 1 k 12
# 2: a 3 2 k 15
# 3: a 6 3 k 18
# 4: b 1 4 k 12
# 5: b 3 5 k 15
# 6: b 6 6 k 18
# 7: c 1 7 k 12
# 8: c 3 8 k 15
# 9: c 6 9 k 18
多指标的多个分类汇总
- 默认汇总名称
DT[,list(mean(v),sum(v)),by=list(x,y)] # keyed by
# x y V1 V2
# 1: a 1 1 1
# 2: a 3 2 2
# 3: a 6 3 3
# 4: b 1 4 4
# 5: b 3 5 5
# 6: b 6 6 6
# 7: c 1 7 7
# 8: c 3 8 8
# 9: c 6 9 9
- 自定义汇总名称
DT[,list(mean.v = mean(v),sum.v = sum(v)),by=list(x,y)] # keyed by
# x y mean.v sum.v
#1: a 1 1 1
#2: a 3 2 2
#3: a 6 3 3
#4: b 1 4 4
#5: b 3 5 5
#6: b 6 6 6
#7: c 1 7 7
#8: c 3 8 8
#9: c 6 9 9
- 汇总结果与原始数据进行匹配
DT[,c("mean.v", "sum.v.y") := list(mean(v),sum(v)) ,by=list(x,y)]
# x y v a sum.v.y mean.v
# 1: a 1 1 k 1 1
# 2: a 3 2 k 2 2
# 3: a 6 3 k 3 3
# 4: b 1 4 k 4 4
# 5: b 3 5 k 5 5
# 6: b 6 6 k 6 6
# 7: c 1 7 k 7 7
# 8: c 3 8 k 8 8
# 9: c 6 9 k 9 9
data.table与data.frame的转化
data.table格式加快了处理速度，而data.frame则更为基础。两者的转化可以通过data.table(),setDT()和setDT()来实现，其中data.table()和setDT()函数可以将数据从data.frame转化为data.table，setDF()函数可以将数据从data.table转化为data.frame。注意使用data.table(),setDT()和setDT()时，参数本身的数据类型也会发生变化。
class(DT)
# [1] "data.table" "data.frame"
class(setDF(DT))
# [1] "data.frame"
class(DT)
# [1] "data.frame"
此外，data.table包还可以与基础包中的duplicated()、unique()、subset() 函数结合使用。 不仅如此，data.table包还有一些基础包的替代函数。rbind()升级版的rbindlist()，可以合并列数不同和列位置不同的数据。比dplyr包中arrange()函数更快的setorder()排序函数。