前言：
學(xué)習(xí)R過(guò)程中，數(shù)據(jù)中往往會(huì)出現(xiàn)一些詭異的符號(hào)，而不是數(shù)值，那么他們到底是什么意思呢？下面帶領(lǐng)大家一探究竟。

(一) 特殊值的概念

1. NaN
R中的無(wú)定義數(shù)用NaN表示，即“Not a Number（非數(shù)）”。
不過(guò)在R中，R實(shí)際上是把NaN視作一個(gè)數(shù)的，當(dāng)其參與運(yùn)算時(shí)，返回結(jié)果總是NaN。我們可以使用is.nan()函數(shù)來(lái)檢測(cè)計(jì)算結(jié)果有無(wú)定義，但是需要注意的是，對(duì)于NaN而言，is.finite()和is.infinite()都會(huì)返回FALSE。

> 0/0
[1] NaN

2. NA
NA表示缺失值，即“Missing value”，是“not available”的縮寫(xiě)

> a <- c(1, 2, 3, 4)
> a[1:5]
[1]  1  2  3  4 NA

3. Inf
R中的無(wú)窮大用Inf表示（即Infinity，無(wú)窮大），負(fù)無(wú)窮表示為-Inf。
要檢查一個(gè)數(shù)是否為無(wú)窮，可以使用is.finite()或者is.infinite()函數(shù)

> 1/0
[1] Inf
> -1/0
[1] -Inf

3. NULL
R語(yǔ)言中，NA代表位置上的值為空，NULL代表連位置都沒(méi)有，變量為空，其長(zhǎng)度為0，表明“空無(wú)一物”

> a <- NULL
> a
NULL
> length(a)
[1] 0

(二) 實(shí)戰(zhàn)中處理無(wú)效數(shù)據(jù)

#設(shè)置sugar數(shù)據(jù)
> sugar <- matrix(1:25,5,5)
> sugar[1,] <- 0
> sugar[3,1] <- 0
> sugar
     [,1] [,2] [,3] [,4] [,5]
[1,]    0    0    0    0    0
[2,]    2    7   12   17   22
[3,]    0    8   13   18   23
[4,]    4    9   14   19   24
[5,]    5   10   15   20   25
#設(shè)置yeast數(shù)據(jù)
> yeast <- matrix(10:34,,5,5)
> yeast[3,] <- 0
> yeast[1,1] <- 0
> yeast
     [,1] [,2] [,3] [,4] [,5]
[1,]    0   11   12   13   14
[2,]   15   16   17   18   19
[3,]    0    0    0    0    0
[4,]   25   26   27   28   29
[5,]   30   31   32   33   34

將兩個(gè)數(shù)據(jù)進(jìn)行除法運(yùn)算

> foldchange <- sugar/yeast
> foldchange
          [,1]      [,2]      [,3]      [,4]      [,5]
[1,]       NaN 0.0000000 0.0000000 0.0000000 0.0000000
[2,] 0.1333333 0.4375000 0.7058824 0.9444444 1.1578947
[3,]       NaN       Inf       Inf       Inf       Inf
[4,] 0.1600000 0.3461538 0.5185185 0.6785714 0.8275862
[5,] 0.1666667 0.3225806 0.4687500 0.6060606 0.7352941
> log2_foldchange <- log2(sugar / yeast)
> log2_foldchange
          [,1]      [,2]       [,3]        [,4]       [,5]
[1,]       NaN      -Inf       -Inf        -Inf       -Inf
[2,] -2.906891 -1.192645 -0.5025003 -0.08246216  0.2115041
[3,]       NaN       Inf        Inf         Inf        Inf
[4,] -2.643856 -1.530515 -0.9475326 -0.55942741 -0.2730185
[5,] -2.584963 -1.632268 -1.0931094 -0.72246602 -0.4436067

而log2_foldchange就是我們需要的數(shù)據(jù)，發(fā)現(xiàn)里面有許多 NaN、 -Inf 、Inf ，想辦法進(jìn)行數(shù)據(jù)清洗。

> yeast == 0 # 邏輯判斷數(shù)據(jù)中是否為0
      [,1]  [,2]  [,3]  [,4]  [,5]
[1,]  TRUE FALSE FALSE FALSE FALSE
[2,] FALSE FALSE FALSE FALSE FALSE
[3,]  TRUE  TRUE  TRUE  TRUE  TRUE
[4,] FALSE FALSE FALSE FALSE FALSE
[5,] FALSE FALSE FALSE FALSE FALSE
> log2_foldchange[yeast == 0] # 選擇里面判斷為0的數(shù)據(jù)
[1] NaN NaN Inf Inf Inf Inf
# 將無(wú)效值替換為0
> log2_foldchange[sugar == 0 | yeast == 0] <- 0
> log2_foldchange
          [,1]      [,2]       [,3]        [,4]       [,5]
[1,]  0.000000  0.000000  0.0000000  0.00000000  0.0000000
[2,] -2.906891 -1.192645 -0.5025003 -0.08246216  0.2115041
[3,]  0.000000  0.000000  0.0000000  0.00000000  0.0000000
[4,] -2.643856 -1.530515 -0.9475326 -0.55942741 -0.2730185
[5,] -2.584963 -1.632268 -1.0931094 -0.72246602 -0.4436067

都到這一步了，再繪制兩個(gè)圖形玩玩:

1. hist圖

> hist(log2_foldchange, col = "red",  border = "black")

2018-10-07Rplot01.png

2. 密度圖

> curve <- density(log2_foldchange)
> plot(curve, main = "understand the NaN Inf NA")
> polygon(curve, col = "Thistle", border = "red", lty =1 )

2018-10-07Rplot02.png