{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# 数据结构\n", "\n", "本节介绍 R 的数据类型，包括 data.frame 相关的重要命令；也给出了转换数据类型的函数。" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 数据类型\n", "\n", "R 中标志的数据类型有（部分在下文详细介绍）：\n", "\n", "- 数字：numeric\n", "- 字符：character\n", "- 向量：vector\n", "- 矩阵：matrix\n", "- 数据框：data.frame\n", "- 因子：factor\n", "- 逻辑：logical\n", "\n", "### 类型判断与转换\n", "\n", "对于上述的每个类型的缩写 *abbr*，对应的判断与转换函数：\n", "\n", "- 判断：is.{abbr}()，例如判断是否是数字 is.numeric()\n", "- 转换：as.{abbr}()" ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "[[1]]\n", "[1] 1 2 3 4\n", "\n", "[[2]]\n", "[1] \"1\" \"2\" \"3\" \"4\"\n", "\n" ] } ], "source": [ "tmp <- c(1:4)\n", "tmp.ch <- as.character(tmp)\n", "print(list(tmp, tmp.ch))" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "此外，还有 class() 与 typeof() 函数可以参考：" ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "[1] \"integer\" \"integer\"\n" ] } ], "source": [ "print(c(class(tmp),\n", " typeof(tmp)))" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### 经典的数据类型\n", "\n", "由一系列数据组成的变量，按照通常数据处理的分类，有：\n", "\n", "- 数值变量（或定量变量，quantitative）：\n", " - 连续型（continuous）：又称定比。比如试卷在班级的及格率，其可以是 $[0,1]$ 之间的任意值。连续型变量内各数据间的大小（或优劣）关系是显然的。\n", " - 离散型（discrete）：又称定距。比如班级的同学个数。这类变量常常是通过计数得到的，其任两个数据之间的差值必定为某基础值的整数倍，如不可能有 16.5 个同学。\n", "- 非数值变量（或定性变量，qualitative）：\n", " - 类别型（categorical）：又称定类。比如同学的主修专业。这类变量中各数据点往往是字符，互相之间无优劣关系。\n", " - 有序型（ordinal）：又称定序。比如成绩的等级，ABCD。互相之间有优劣顺序。\n", "\n", "对于一般的二维数据集，其行与列在不同领域的称呼不同：\n", "\n", "| 领域 | 行 | 列 |\n", "| --- | --- | --- |\n", "| 统计学（本文） | 观测(observation) | 变量(variable) |\n", "| 数据库 | 记录(record) | 字段(field) |\n", "| 数据挖掘/机器学习 | 示例(example) | 属性(attribute) |\n", "\n", "## 因子(factor)\n", "\n", "在 R 中，将非数值变量统称为**因子**（factor），并分为有序因子与无序因子两种。例如，我们有主修专业数据：" ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "[1] \"Engineering\" \"Math\" \"Social Science\"\n" ] } ], "source": [ "prog <- c(\"Math\", \"Engineering\", \"Social Science\", \"Math\")\n", "tmp <- factor(prog)\n", "\n", "print(levels(tmp)) # 返回因子的类别" ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ " Factor w/ 3 levels \"Engineering\",..: 2 1 3 2\n" ] } ], "source": [ "str(tmp) # 显示因子类别数。str 是 structure 的缩写" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "从上面的 str(tmp) 中可以看到，Level 1对应 Engineering，2对应Math，3对应Social Science。因为这是根据字符串首字母顺序编号的。\n", "\n", "如果你想从数值记录的变量生成一个因子，使用参数 levels 与 labels ：" ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ " Factor w/ 3 levels \"Math\",\"Engineering\",..: 1 2 3 1\n" ] } ], "source": [ "tmp <- factor(c(1, 2, 3, 1), levels=c(1,2,3), \n", " labels=c(\"Math\", \"Engineering\", \"Social Science\"))\n", "str(tmp)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "如果想把有序型变量转成因子，并指定顺序，可以使用参数 order= ：" ] }, { "cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ " Ord.factor w/ 3 levels \"B\"<\"A\"<\"C\": 2 1 3 1 3 2\n" ] } ], "source": [ "tmp <- factor(c(\"A\", \"B\", \"C\", \"B\", \"C\", \"A\"), order=TRUE,\n", " levels=c(\"B\", \"A\", \"C\"))\n", "str(tmp)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "还有一点需要注意：在我们对数据框进行了合并、筛选等处理后，**因子可能不再含有全部水平对应的值**（比如原有3个水平，但其中一个水平的值都被筛除了）。这时候需要使用 droplevels() 命令来抛弃多余的水平：\n", "\n", "```r\n", "dt[\"factor\"] <- droplevels(dt[\"dactor\"])\n", "```" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 同类元素集：向量、矩阵与数组\n", "\n", "**同类型元素**的一维堆叠叫向量，二维堆叠叫矩阵。更高维的叫数组。\n", "\n", "### 向量\n", "\n", "如果向量只有一个元素，称为标量。" ] }, { "cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "[1] 1 2 3\n" ] } ], "source": [ "vector <- c(1:3)\n", "print(vector)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### 矩阵\n", "\n", "矩阵的建立方法就复杂一些：" ] }, { "cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ " [,1] [,2] [,3] [,4]\n", "[1,] 1 2 3 4\n", "[2,] 5 6 7 8\n", "[3,] 9 10 11 12\n" ] } ], "source": [ "# matr <- matrix(d, nrow=N, ncol=N, byrow=T/F, [dimnames=list(rownamestr,colnamestr)])\n", "\n", "matr <- matrix(1:12, nrow=3, ncol=4, byrow=T)\n", "print(matr)" ] }, { "cell_type": "code", "execution_count": 9, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ " 1 2 3 4\n", "A 1 4 7 10\n", "B 2 5 8 11\n", "C 3 6 9 12\n" ] } ], "source": [ "# 按列填充，并加上行名与列名\n", "rowname <- c('A', 'B', 'C')\n", "colname <- c(as.character(1:4)) # 转为字符串\n", "matr <- matrix(1:12, nrow=3, ncol=4, byrow=F, dimnames=list(rowname, colname))\n", "print(matr)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "下标的使用很简单。" ] }, { "cell_type": "code", "execution_count": 10, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ " 1 2 3 4 \n", " 2 5 8 11 \n" ] } ], "source": [ "print(matr[2,]) # 第二行" ] }, { "cell_type": "code", "execution_count": 11, "metadata": {}, "outputs": [ { "data": { "text/html": [ "\n", "\n", "\n", "\t\n", "\t\n", "\n", "
13
B28
C39
\n" ], "text/latex": [ "\\begin{tabular}{r|ll}\n", " & 1 & 3\\\\\n", "\\hline\n", "\tB & 2 & 8\\\\\n", "\tC & 3 & 9\\\\\n", "\\end{tabular}\n" ], "text/markdown": [ "\n", "| | 1 | 3 | \n", "|---|---|\n", "| B | 2 | 8 | \n", "| C | 3 | 9 | \n", "\n", "\n" ], "text/plain": [ " 1 3\n", "B 2 8\n", "C 3 9" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "matr[2:3,c(1,3)] # 第2、3行，第1、3列" ] }, { "cell_type": "code", "execution_count": 12, "metadata": {}, "outputs": [ { "data": { "text/html": [ "\n", "\n", "\n", "\t\n", "\t\n", "\t\n", "\n", "
34
A7 10
B8 11
C9 12
\n" ], "text/latex": [ "\\begin{tabular}{r|ll}\n", " & 3 & 4\\\\\n", "\\hline\n", "\tA & 7 & 10\\\\\n", "\tB & 8 & 11\\\\\n", "\tC & 9 & 12\\\\\n", "\\end{tabular}\n" ], "text/markdown": [ "\n", "| | 3 | 4 | \n", "|---|---|---|\n", "| A | 7 | 10 | \n", "| B | 8 | 11 | \n", "| C | 9 | 12 | \n", "\n", "\n" ], "text/plain": [ " 3 4 \n", "A 7 10\n", "B 8 11\n", "C 9 12" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "matr[,-c(1:2)] # 除前2列外所有列" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "几个注意点：\n", "\n", "- 矩阵的对应元素乘法是“\\*”。如果要使用线性代数中的**矩阵乘法**，使用 “%\\*%”。\n", "- 矩阵的转置：t(matr)\n", "- 矩阵的逆：solve(matr)。该函数 solve(a, b) 是用于求解线性方程组 aX=b 的，b 缺省时求逆。" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### 数组\n", "\n", "array() 语法类似于 matrix() 。" ] }, { "cell_type": "code", "execution_count": 13, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ ", , C1\n", "\n", " B1 B2 B3\n", "A1 1 3 5\n", "A2 2 4 6\n", "\n", ", , C2\n", "\n", " B1 B2 B3\n", "A1 7 9 11\n", "A2 8 10 12\n", "\n", ", , C3\n", "\n", " B1 B2 B3\n", "A1 13 15 17\n", "A2 14 16 18\n", "\n", ", , C4\n", "\n", " B1 B2 B3\n", "A1 19 21 23\n", "A2 20 22 24\n", "\n" ] } ], "source": [ "# arr <- array(vector, dimensions, dimnames)\n", "# dimensions: 指定数组有几维，以及各维的长度\n", "\n", "arr <- array(1:24, c(2, 3, 4), \n", " dimnames=list(c('A1', 'A2'), c('B1', 'B2','B3'), c('C1', 'C2', 'C3', 'C4')))\n", "print(arr)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 列表\n", "\n", "列表是一系列变量的有序组合。声明方式：\n", "\n", " lst <- list(name1=obj1, name2=obj2, ...)" ] }, { "cell_type": "code", "execution_count": 14, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "$prog\n", "[1] \"Math\" \"Engineering\"\n", "\n", "$gender\n", "[1] 1 2\n", "\n", "$grade\n", "[1] \"No grade\"\n", "\n" ] } ], "source": [ "lst <- list(prog=c(\"Math\", \"Engineering\"), gender=c(1, 2), grade=\"No grade\")\n", "print(lst)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "你可以通过美元符来调用它们，如：" ] }, { "cell_type": "code", "execution_count": 15, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "[1] \"Math\" \"Engineering\"\n" ] } ], "source": [ "print(lst$prog)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## 数据框：data.frame\n", "\n", "数据框是多个等长向量的按列堆叠。这是 R 中最常用的数据结构。由于此内容的重要性，我将其提升了一级大纲。" ] }, { "cell_type": "code", "execution_count": 16, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ " name age gender\n", "1 Allen 20 Male\n", "2 Bruth 21 Male\n", "3 Chris 22 Male\n", "4 Daisy 23 Female\n" ] } ], "source": [ "# df <- data.frame(, [row.names=])\n", "\n", "name <- c('Allen', 'Bruth', 'Chris', 'Daisy')\n", "age <- c(20:23)\n", "gender <- c('Male', 'Male', 'Male', 'Female')\n", "\n", "df <- data.frame(name, age, gender)\n", "print(df)" ] }, { "cell_type": "code", "execution_count": 17, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ " name age gender\n", "Allen Allen 20 Male\n", "Bruth Bruth 21 Male\n", "Chris Chris 22 Male\n", "Daisy Daisy 23 Female\n" ] } ], "source": [ "# 指定唯一标识的一列作为行名\n", "df <- data.frame(name, age, gender, row.names=name)\n", "print(df)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "数据框默认按照列进行选取：" ] }, { "cell_type": "code", "execution_count": 18, "metadata": {}, "outputs": [ { "data": { "text/html": [ "\n", "\n", "\n", "\t\n", "\t\n", "\t\n", "\t\n", "\n", "
namegender
AllenAllen Male
BruthBruth Male
ChrisChris Male
DaisyDaisy Female
\n" ], "text/latex": [ "\\begin{tabular}{r|ll}\n", " & name & gender\\\\\n", "\\hline\n", "\tAllen & Allen & Male \\\\\n", "\tBruth & Bruth & Male \\\\\n", "\tChris & Chris & Male \\\\\n", "\tDaisy & Daisy & Female\\\\\n", "\\end{tabular}\n" ], "text/markdown": [ "\n", "| | name | gender | \n", "|---|---|---|---|\n", "| Allen | Allen | Male | \n", "| Bruth | Bruth | Male | \n", "| Chris | Chris | Male | \n", "| Daisy | Daisy | Female | \n", "\n", "\n" ], "text/plain": [ " name gender\n", "Allen Allen Male \n", "Bruth Bruth Male \n", "Chris Chris Male \n", "Daisy Daisy Female" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "df[c(1,3)]" ] }, { "cell_type": "code", "execution_count": 19, "metadata": {}, "outputs": [ { "data": { "text/html": [ "\n", "\n", "\n", "\t\n", "\t\n", "\t\n", "\t\n", "\n", "
namegender
AllenAllen Male
BruthBruth Male
ChrisChris Male
DaisyDaisy Female
\n" ], "text/latex": [ "\\begin{tabular}{r|ll}\n", " & name & gender\\\\\n", "\\hline\n", "\tAllen & Allen & Male \\\\\n", "\tBruth & Bruth & Male \\\\\n", "\tChris & Chris & Male \\\\\n", "\tDaisy & Daisy & Female\\\\\n", "\\end{tabular}\n" ], "text/markdown": [ "\n", "| | name | gender | \n", "|---|---|---|---|\n", "| Allen | Allen | Male | \n", "| Bruth | Bruth | Male | \n", "| Chris | Chris | Male | \n", "| Daisy | Daisy | Female | \n", "\n", "\n" ], "text/plain": [ " name gender\n", "Allen Allen Male \n", "Bruth Bruth Male \n", "Chris Chris Male \n", "Daisy Daisy Female" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# 可以按列名选取\n", "df[c(\"name\", \"gender\")]" ] }, { "cell_type": "code", "execution_count": 20, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "[1] 20 21 22 23\n" ] } ], "source": [ "# 可以通过标识符 '$' 选取\n", "print(df$age)" ] }, { "cell_type": "code", "execution_count": 21, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "[1] Male Male Male\n", "Levels: Female Male\n" ] } ], "source": [ "# 双参数时，先行后列\n", "print(df[1:3,\"gender\"])" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### 临时环境：attach() 与 with()\n", "\n", "df\\$age 从语法上说很清晰，但是可读性却不高。因为我们一般总是在处理一个数据集，因此‘df’显得多余。这里可以借助 attach()/detach() 命令组：\n", "\n", "*请注意：此命令出于容错性、可读性考虑，在任何情况下均__不建议__使用。*" ] }, { "cell_type": "code", "execution_count": 22, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "[1] 20 21 22 23\n" ] } ], "source": [ "# 删除原有的全局变量，否则 age 会覆盖 df$age\n", "rm(age, gender, name)\n", "\n", "attach(df)\n", "print(age)\n", "detach(df)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "或者使用 with() 命令：" ] }, { "cell_type": "code", "execution_count": 23, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "[1] Allen Bruth Chris Daisy\n", "Levels: Allen Bruth Chris Daisy\n", "[1] 2\n" ] } ], "source": [ "tmp <- 1\n", "with(df, {\n", " print(name)\n", " tmp <<- 2 # 特殊赋值符可以传值到 with() 之外\n", "})\n", "print(tmp)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### 更多内容\n", "\n", "更多关于数据管理和 data.frame 处理的内容在本系列文章的 [数据管理](ManageData.ipynb) 一文中可以找到。" ] } ], "metadata": { "kernelspec": { "display_name": "R", "language": "R", "name": "ir" }, "language_info": { "codemirror_mode": "r", "file_extension": ".r", "mimetype": "text/x-r-source", "name": "R", "pygments_lexer": "r", "version": "4.0.0" }, "toc": { "base_numbering": 1, "nav_menu": {}, "number_sections": true, "sideBar": true, "skip_h1_title": false, "title_cell": "Table of Contents", "title_sidebar": "Contents", "toc_cell": false, "toc_position": {}, "toc_section_display": true, "toc_window_display": false } }, "nbformat": 4, "nbformat_minor": 2 }