# base 函数 ## 赋值运算符 `<-` | Left assignment |`->` | Right assignment |`<<-` | Left lexicographic assignment (for advanced users) |`->>` | Right lexicographic assignment (for advanced users) |## 其他运算符 ## 逻辑索引 ```{r} 1 : length (mtcars$ mpg)seq_along (mtcars$ mpg)<- c (1 ,3 ,6 ,4 ,0 ,7 )# 返回值为TRUE的索引 which (x> 2 )which.max (x)which.min (x)# 返回值 > 2 ]which (x> 2 )]# 包含 # match(x, table, nomatch = NA_integer_, incomparables = NULL) # x %in% table 1 : 10 %in% c (1 ,3 ,5 ,9 )match (1 : 10 , c (1 ,3 ,5 ,9 ))all (colnames (mpg)== colnames (mpg))any (colnames (mpg)== colnames (mtcars))<- c (9 : 20 , 1 : 5 , 3 : 7 , 0 : 8 )duplicated (x)<- x[! duplicated (x)]``` ## 子集运算 `[` ，`$` ，`[[` `@` `slot` ### `[ ]` ```{r} <- c (2.1 , 4.2 , 3.3 , 5.4 )c (3 , 1 )]- c (3 , 1 )]c (TRUE , TRUE , FALSE , FALSE )]> 3 ]c (TRUE , FALSE )] # recycling rules 循环 0 ]``` ### `$`, `[[ ]]` `$` 是一个简写运算符, `x$y` 大致相当于 `x[["y"]]` ,从左到右部分赋值```{r} <- list (abc = 1 )$ a"a" ]]options (warnPartialMatchDollar = TRUE )$ a``` ## 生成数值序列 ```{r} seq (from= 1 ,to= 30 ,by= 3 ) # 生成一个序列 seq (from= 1 ,to= 30 ,length= 10 ) sequence (30 ,by= 3 )rep (x = c ("A" ,1 ,"B" ,2 ),times = 3 ) #重复序列 rep (x = c ("A" ,1 ,"B" ,2 ),times = c (1 ,2 ,3 ,4 ))rep (x = c ("A" ,1 ,"B" ,2 ),each = 3 ,times = 2 )``` ## 排序 ```{r} <- c ("b" , "c" , "a" )order (x)order (x)]``` ```{r eval=FALSE} sort() dplyr::arrange() # 分配秩次 rank() ``` ## 拆分 ```{r} set.seed (3 )<- CO2[sample (1 : nrow (CO2), 10 ), ]head (df)<- split (df, f = list (df$ Type, df$ Treatment))``` ## 合并 ```{r} #| code-fold: true set.seed (61 )<- 1 : 10 <- c ("Andrew" , "Susan" , "John" , "Joe" , "Jack" ,"Jacob" , "Mary" , "Kate" , "Jacqueline" , "Ivy" )<- round (rnorm (10 , mean = 1500 , sd = 200 ))<- round (rnorm (10 , mean = 50 , sd = 8 ))<- c ("CTO" , "CFO" , "Administrative" , rep ("Technician" , 7 ))<- data.frame (id = employee_id[1 : 8 ],name = employee_name[1 : 8 ],month_salary = employee_salary[1 : 8 ])<- data.frame (id = employee_id[- 5 ],name = employee_name[- 5 ],age = employee_age[- 5 ],position = employee_position[- 5 ]``` ### 连接 ```{r} # 内连接 merge (x = df_1, y = df_2, by = c ("id" , "name" )) # 全连接 merge (x = df_1, y = df_2, all = TRUE )# 左连接 merge (x = df_1, y = df_2, all.x = TRUE )# 右连接 merge (x = df_1, y = df_2, all.y = TRUE )# 交叉连接 <- merge (x = df_1, y = df_2, by = NULL )head (Merged)``` ### 按行名合并 ```{r} <- data.frame (var = c ("one" , "two" , "three" , "four" , "five" ),data = c (1 , 5 , 1 , 6 , 8 ))rownames (df1) <- c ("A" , "B" , "C" , "D" , "E" )<- data.frame (var = c ("three" , "one" , "eight" , "two" , "nine" ),data = c (1 , 5 , 1 , 6 , 8 ))rownames (df2) <- c ("E" , "A" , "B" , "D" , "C" )``` ```{r} merge (df1, df2, by = 0 , all = TRUE ) merge (df1, df2, by = "row.names" , all = TRUE ) # Equivalent ``` ## AsIs `I()` `I()` 函数来自于`base` 包，它是as.is函数的别名。I()函数用于改变对象的类别，指示R在进行操作时应该将该对象视为其原始形式，不对它进行任何转换。```{r} lapply (summary (mtcars$ mpg), I) lapply (summary (mtcars$ mpg), I) |> as_tibble ()``` ## 数学函数 ```{r} <- c (25 ,- 4 ,3.66 ,3.42 ,- 5.99 )abs (x) #绝对值 sqrt (x) #平方根 ceiling (x) #向上取整 floor (x) #向下取整 trunc (x) #整数部分 round (x,digits = 2 ) #四舍五入，保留2位小数 signif (x,digits = 2 ) #四舍五入，保留2有效数字 log (x,base= 4 ) #对x取以base为底的对数 log (x) #自然对数 log10 (x) #常用对数 exp (x) #e指数函数 ``` ## 统计函数 ```{r} mean (x,trim = 0.05 ,na.rm = TRUE ) #算术平均值 median (x) #中位数 sd (x) #标准差 var (x) #方差 mad (x) #绝对中位差 quantile (x,probs = c (0 ,0.25 ,0.5 ,0.75 ,1 )) #分位数 range (x) # 值域 scale (x,center = TRUE ,scale = TRUE ) #标准化（均值为0、标准差为1） scale (x,center = TRUE ,scale = FALSE ) #中心化：减去均值 <- c (1 ,3 ,5 ,7 ,9 ,11 ,13 )# Find the "previous" (lag()) or "next" (lead()) values in a vector :: lag (x,n= 2 ) # n阶滞后 :: lead (x,n= 2 ) # n阶前移 # 滞后差分 lag阶滞后 difference阶差分 diff (x,lag = 1 ,difference= 1 ) # 隔0个值后位减前位，进行1次 diff (x,lag = 1 ,difference= 2 ) ## 隔0个值后位减前位，进行2次 diff (x,lag = 1 ,difference= 3 ) # 隔0个值后位减前位，进行3次 diff (x,lag = 2 ,difference= 1 ) # 隔1个值后位减前位，进行1次 diff (x,lag = 2 ,difference= 2 ) ## 隔1个值后位减前位，进行2次 ``` ## 字符串函数 ```{r} # 匹配 <- data.frame (name = c ("Alice" , "Bob" , "Carol" , "Dave" , "Eve" ),description = c ("Software developer" , "Data analyst" , "UX designer" , "Project manager" , "Data scientist" )$ has_data_analyst <- str_detect (data$ description, "Data analyst" )print (data)$ has_data_grepl <- grepl ("Data" , data$ description)print (data)``` ```{r} #子串 substr (x= "qwertyyuio" ,start = 2 ,stop= 4 ) str_sub (string = "qwertyyuio" ,start = 2 ,end = 4 )#查找替换 sub (pattern = " " ,replacement = "." ,x= "hello world hello !" ,ignore.case = FALSE ,fixed = FALSE ) str_replace_all ("hello world hello !" ," " ,replacement = "." )# 查找,返回下标 grep (pattern = "v" ,x= c ("a" ,"v" ,"D" ,"A" ,"f" ,"J" ),ignore.case = FALSE ,fixed = FALSE ) # 分隔，\\转义字符 strsplit (x= "a.fa.fag" ,split = " \\ ." ,fixed = FALSE ) #连接 paste ("x" ,c ("a" ,"b" ),sep= "" ,collapse = "?" ) paste0 ("x" ,c ("A" ,"B" ),collapse= "?" )cat ("hello" ,"BOb" ," \b\n " ," \b Isn \' R" ," \t " ,"GREAT? \n " ,sep = " " )# toupper ("abc" ) #大写转换 tolower ("aaAGEErg" ) #小写转换 ``` ## 打印函数 `sprintf()` ```{r} <- "string" sprintf ("This is where a %s goes." , a)<- 8 sprintf ("Regular:%04d" , x)sprintf ("%f" , pi) # "3.141593" sprintf ("%.3f" , pi) # "3.142" sprintf ("%1.0f" , pi) # "3" sprintf ("%5.1f" , pi) # " 3.1" sprintf ("%05.1f" , pi) # "003.1" sprintf ("%+f" , pi) # "+3.141593" sprintf ("% f" , pi) # " 3.141593" sprintf ("%-10f" , pi) # "3.141593 " (left justified) sprintf ("%e" , pi) #"3.141593e+00" sprintf ("%E" , pi) # "3.141593E+00" sprintf ("%g" , pi) # "3.14159" sprintf ("%g" , 1e6 * pi) # "3.14159e+06" (exponential) sprintf ("%.9g" , 1e6 * pi) # "3141592.65" ("fixed") sprintf ("%G" , 1e-6 * pi) # "3.14159E-06" <- "string" sprintf ("Substitute in multiple strings: %s %s" , x, "string2" )sprintf ("A single percent sign here： %%" )``` ## apply函数簇 ### `apply()` ```{r} apply (X = mtcars, MARGIN = 2 , FUN = mean)``` ### `aggregate()` ```{r} aggregate (x = mtcars,by = list (am= mtcars$ am,cyl= mtcars$ cyl),FUN = mean) aggregate (.~ am+ cyl,data = mtcars,FUN = mean)aggregate (mpg~ am+ cyl,data = mtcars,FUN = summary)``` ### `by()`, `tapply()` `by()` 是应用于数据框的tapply的面向对象包装器。```{r} by (data = mtcars[,1 : 3 ],INDICES = list (cyl= mtcars$ cyl),FUN = summary)by (data = mtcars[,1 : 3 ],INDICES = factor (mtcars$ cyl),FUN = function (x) lm ( disp ~ mpg, data = x))tapply (X = mtcars[,1 : 3 ],INDEX = list (cyl= mtcars$ cyl),FUN = summary)``` ### `lapply()`, `sapply()` `lapply()` returns a list of the same length as X,`sapply()` is a user-friendly version and wrapper of `lapply()` by default returning a vector or matrix```{r} <- list (a = 1 : 10 , beta = exp (- 3 : 3 ), logic = c (TRUE ,FALSE ,FALSE ,TRUE ))lapply (X = x,FUN = quantile)sapply (x, quantile)``` ### `vapply` ```{r} <- split (mtcars, mtcars$ cyl)<- lapply (by_cyl, function (data) lm (mpg ~ wt, data = data))vapply (models, function (x) coef (x)[[2 ]], double (1 ))``` ## 矩阵运算 ```{r} <- matrix (data = 1 : 9 ,ncol = 3 ,nrow = 3 )diag (X)det (X)eigen (X)# 协方差矩阵 cov (X)# 相关系数矩阵 cor (X)``` ## 统计分布函数 `[dpqr]distribution_abbreviation` ，其中密度函数`d` ，分布函数`p` ，分位数函数`q` ，随机数生成函数`r` ```{r} ### 正态分布 dnorm (3 ,0 ,2 ) #N(0,4)在 3 处的密度值 pnorm (1 : 3 ,0 ,2 )#N(0,4)在1，2，3处的分布概率值 qnorm (1 -0.025 ,0 ,1 )# N(0,1)的上0.025分位数 rnorm (5 ,3 ,3 ) # 生成5个服从N(3,9)的随机数 ### 二项分布 1 - sum (dbinom (0 : 1 ,400 ,0.02 ))# 命中率为0.02，独立射击400次，至少击中2次的概率 ggplot ()### 多元正态分布 <- c (230.7 ,146.7 ,3.6 )<- matrix (c (15360.8 ,6721.2 ,- 47.1 ,6721.2 ,4700.9 ,- 16.5 ,- 47.1 ,16.5 ,0.3 ),nrow = 3 ,ncol = 3 )library (MASS)<- mvrnorm (500 ,mean,sigma)head (multi)``` 