R語言基礎(chǔ)之第三部分：重要函數(shù)apply族函數(shù)的使用
3.1 apply族函數(shù)
3.11 lapply
可以看成是list + apply的縮寫
循環(huán)處理列表中的每一個(gè)元素，總是返回一個(gè)列表
lapply（列表，函數(shù)/函數(shù)名，其他參數(shù)）
> str(lapply) #查看函數(shù)怎么用吧
> function (X, FUN, ...)
> x = list(a = 1:10, b = c(21,31,41,51,61))
> lapply(x, mean) #分別作用列表x中的a元素 和 b元素
$a
[1] 5.5
$b
[1] 41
----------------------------------我是美麗的短暫分割線------------------------------
> y = 1: 4
> lapply(y,runif) # 注意：1:4 表示的列表的內(nèi)容，而不是列表名稱
# 該函數(shù)自動(dòng)把y這個(gè)向量格式轉(zhuǎn)換為了列表
[[1]]
[1] 0.9995728
[[2]]
[1] 0.7146108 0.9089188
[[3]]
[1] 0.9894942 0.5711674 0.6793558
[[4]]
[1] 0.8209915 0.6317327 0.3662067 0.5019538
> lapply(y,runif,min = 0, max = 100)
[[1]]
[1] 82.04034
[[2]]
[1] 57.51863 32.18876
[[3]]
[1] 65.578491 63.255324 8.271222
[[4]]
[1] 65.20686 30.99439 13.01915 72.22398
----------------------------------我是美麗的短暫分割線------------------------------
> z = list(a = matrix(1:6,2,3), b = matrix(4:7,2,2))
> z
$a
[,1] [,2] [,3]
[1,] 1 3 5
[2,] 2 4 6
$b
[,1] [,2]
[1,] 4 6
[2,] 5 7
> lapply(z, function(m) m[1,]) # **因?yàn)閦的每個(gè)元素都為矩陣，所以在此m也就表示默認(rèn)矩陣**
$a
[1] 1 3 5
$b
[1] 4 6
3.12 sapply
與lapply相似，唯一的區(qū)別是：輸出的結(jié)果能簡潔就簡潔
> x = list(a = 1:10, b =c(11,21,31,41,51))
> lapply(x,mean)
$a
[1] 5.5
$b
[1] 31
> sapply(x,mean)
a b
5.5 31.0
規(guī)則就是：當(dāng)結(jié)果列表的元素長度均為1時(shí)，返回向量；
當(dāng)結(jié)果列表的元素長度相同，且大于1，返回矩陣；
若沒有列表元素雜七雜八，返回列表，即lapply結(jié)果。
3.13 apply
沿著數(shù)組的某一維度處理數(shù)據(jù)
apply（數(shù)組，維度，函數(shù)）
> x = matrix(1:16, 4, 4)
> x
[,1] [,2] [,3] [,4]
[1,] 1 5 9 13
[2,] 2 6 10 14
[3,] 3 7 11 15
[4,] 4 8 12 16
> apply(x,1,mean) #計(jì)算每行的均值1表示行 2表示列
[1] 7 8 9 10
# 求列均值 列和 行均值 行和有簡單函數(shù)
> colMeans(x)
> colSums(x)
> rowMeans(x)
> rowSums(x)
----------------------------------我是美麗的短暫分割線------------------------------
> x = matrix(rnorm(100),10,10) #生成100個(gè)標(biāo)態(tài)分布的隨機(jī)數(shù)并組成10行10列矩陣
> apply(x,1,quantile,probs = c(0.25,0.75)) #求每行25% 和75% 的分位數(shù)
[,1] [,2] [,3] [,4] [,5] [,6] [,7]
25% -0.662055 0.08879356 -0.1460079 -0.6126172 -0.38371665 -0.2321496 -0.9739394
75% 1.546184 0.81950212 1.1570631 0.5038655 0.06949848 1.7607837 0.4571585
[,8] [,9] [,10]
25% -0.4376032 -0.5294471 -0.5692924
75% 0.8542543 1.2788572 0.4973777
----------------------------------我是美麗的短暫分割線------------------------------
z = array(rnorm(2*3*4),c(2,3,4))
> z
, , 1
[,1] [,2] [,3]
[1,] 0.90596834 1.0847827 -0.8454799
[2,] 0.08787108 -0.6741413 -1.2779192
, , 2
[,1] [,2] [,3]
[1,] 0.7279739 1.0178221 -0.9207198
[2,] -1.0621578 0.6427292 -1.5874224
, , 3
[,1] [,2] [,3]
[1,] -1.2396928 0.6654546 -0.9154209
[2,] -0.9939213 1.1634397 1.4078552
, , 4
[,1] [,2] [,3]
[1,] -0.1845830 -1.7680792 -0.47874217
[2,] 0.4592537 -0.4854505 -0.06110676
apply(z,c(1,2),mean)
[,1] [,2] [,3]
[1,] 0.05241661 0.2499950 -0.7900907
[2,] -0.37723859 0.1616443 -0.3796483
> apply(z,c(1,3),mean)
[,1] [,2] [,3] [,4]
[1,] 0.3817571 0.2750254 -0.4965530 -0.81046815
[2,] -0.6213965 -0.6689503 0.5257912 -0.02910119
> apply(z,c(2,3),mean)
[,1] [,2] [,3] [,4]
[1,] 0.4969197 -0.1670919 -1.1168070 0.1373353
[2,] 0.2053207 0.8302756 0.9144472 -1.1267649
[3,] -1.0616995 -1.2540711 0.2462172 -0.2699245
3.14 mapply
lapply的多元版本
mapply（函數(shù)/函數(shù)名，數(shù)據(jù)，函數(shù)相關(guān)的函數(shù)）
> list(rep(1,4), rep(2,3),rep(3,2),rep(4,1))
[[1]]
[1] 1 1 1 1
[[2]]
[1] 2 2 2
[[3]]
[1] 3 3
[[4]]
[1] 4
> mapply(rep,1:4,4:1)
[[1]]
[1] 1 1 1 1
[[2]]
[1] 2 2 2
[[3]]
[1] 3 3
[[4]]
[1] 4
> s = function(n, mean, std){
+ rnorm(n, mean, std) }
> s(4,0,1)
[1] 1.1924366 1.6924980 0.1336761 0.4412759
> mapply(s, 1:5, 5:1, 2) #n表示1:5 mean表示5:1 std表示2
[[1]]
[1] 5.79772
[[2]]
[1] 2.018453 5.740265
[[3]]
[1] 3.051907 -2.432438 5.262420
[[4]]
[1] 2.2291033 3.6701801 1.3136980 -0.5424045
[[5]]
[1] 4.3898086 -3.9646175 -0.6947878 1.0125224 3.6488662
3.15 tapply
對(duì)向量子集操作
tapply（向量，因子/因子列表，函數(shù)/函數(shù)名）
> x = c(rnorm(5), runif(5), rnorm(5,1)) # rnorm(5,1) 表示生成5個(gè)均值為1 方差為0的隨機(jī)數(shù)
> f = gl(3,5) #**該因子有3個(gè)水平，每個(gè)水平下有5個(gè)元素**
> f
[1] 1 1 1 1 1 2 2 2 2 2 3 3 3 3 3
Levels: 1 2 3
> tapply(x, f, mean)
1 2 3
0.2099737 0.3680760 1.2042982
> tapply(x, f, mean, simplify = FALSE)
$`1`
[1] 0.2099737
$`2`
[1] 0.368076
$`3`
[1] 1.204298
3.16 split
根據(jù)因子or因子列表將向量or其他對(duì)象分組
通常與lapply函數(shù)一起使用
split（向量/列表/數(shù)據(jù)框，因子/因子列表）
> x = c(rnorm(5), runif(5), rnorm(5,1))
> f = gl(3,5)
> split(x,f)
$`1`
[1] 0.1417127 -0.2330801 -1.0763036 -0.4488686 2.6664082
$`2`
[1] 0.4813495 0.8214398 0.1087439 0.2505093 0.1783375
$`3`
[1] 0.1994114 1.2881557 0.8560102 1.7501790 1.9277348
> lapply(split(x,f), mean)
$`1`
[1] 0.2099737
$`2`
[1] 0.368076
$`3`
[1] 1.204298
> s = split(airquality, airquality$Month) #將airquality按照月分類
> table(airquality$Month)
5 6 7 8 9
31 30 31 31 30
> lapply(s,function(x) colMeans(x[, c("Ozone","Wind","Temp")])) #x表示數(shù)據(jù)框，因?yàn)閟本身是數(shù)據(jù)框。對(duì)s的三個(gè)變量求列均值
$`5`
Ozone Wind Temp
NA 11.62258 65.54839
$`6`
Ozone Wind Temp
NA 10.26667 79.10000
$`7`
Ozone Wind Temp
NA 8.941935 83.903226
$`8`
Ozone Wind Temp
NA 8.793548 83.967742
$`9`
Ozone Wind Temp
NA 10.18 76.90
> sapply(s,function(x) colMeans(x[,c("Ozone","Wind","Temp")],na.rm = TRUE)) #這樣na.rm后結(jié)果中不會(huì)有缺失值
5 6 7 8 9
Ozone 23.61538 29.44444 59.115385 59.961538 31.44828
Wind 11.62258 10.26667 8.941935 8.793548 10.18000
Temp 65.54839 79.10000 83.903226 83.967742 76.90000
后續(xù)請(qǐng)參考：
R語言基礎(chǔ)之第一部分 : 5種數(shù)據(jù)對(duì)象類型
R語言基礎(chǔ)之第二部分 : 操縱數(shù)據(jù)取子集
R語言基礎(chǔ)之第四部分 : 排序
R語言基礎(chǔ)之第五部分 : 總結(jié)數(shù)據(jù)信息
