processing/data_mapping.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "reference : https://stackoverflow.com/questions/16992713/translate-every-element-in-numpy-array-according-to-key"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import seaborn as sns\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "pd.options.display.max_rows = 6"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>diet</th>\n",
       "      <th>kind</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>low fat</td>\n",
       "      <td>rest</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>low fat</td>\n",
       "      <td>rest</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>low fat</td>\n",
       "      <td>rest</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2</td>\n",
       "      <td>low fat</td>\n",
       "      <td>rest</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2</td>\n",
       "      <td>low fat</td>\n",
       "      <td>rest</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   id     diet  kind\n",
       "0   1  low fat  rest\n",
       "1   1  low fat  rest\n",
       "2   1  low fat  rest\n",
       "3   2  low fat  rest\n",
       "4   2  low fat  rest"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = sns.load_dataset('exercise').iloc[:,[1,2,5]]\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['rest', 'walking', 'running'], dtype='object')"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['kind'].cat.categories"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- [일반적인 데이터 mapping](#일반적인-데이터-mapping)\n",
    "\n",
    "- [데이터 조건에 따른 mapping](#데이터-조건에-따른-mapping)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 일반적인 데이터 mapping"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    ".replace 함수 : 지정된 값만 찾아 변경"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "map_dict= {'rest': 'first' , 'walking': 'second', 'running': 'third'}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0       first\n",
       "1       first\n",
       "2       first\n",
       "       ...   \n",
       "87    running\n",
       "88    running\n",
       "89    running\n",
       "Name: kind, Length: 90, dtype: object"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['kind'].replace(map_dict)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### .map 함수 활용"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "replace 함수도 있지만 데이터가 많을수록 .map을 활용하는 것이 더 빠르다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0     first\n",
       "1     first\n",
       "2     first\n",
       "      ...  \n",
       "87      NaN\n",
       "88      NaN\n",
       "89      NaN\n",
       "Name: kind, Length: 90, dtype: object"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['kind'].map(map_dict)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "단, dict에 지정하지 않은 값들은 Nan으로 처리된다. // numpy 방법은 해결 가능"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "map_dict= {'rest': 'first' , 'walking': 'second'}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0     first\n",
       "1     first\n",
       "2     first\n",
       "      ...  \n",
       "87      NaN\n",
       "88      NaN\n",
       "89      NaN\n",
       "Name: kind, Length: 90, dtype: object"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['kind'].map(map_dict)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### numpy 활용"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['first', 'first', 'first', 'first', 'first', 'first', 'first',\n",
       "       'first', 'first', 'first', 'first', 'first', 'first', 'first',\n",
       "       'first', 'first', 'first', 'first', 'first', 'first', 'first',\n",
       "       'first', 'first', 'first', 'first', 'first', 'first', 'first',\n",
       "       'first', 'first', 'second', 'second', 'second', 'second', 'second',\n",
       "       'second', 'second', 'second', 'second', 'second', 'second',\n",
       "       'second', 'second', 'second', 'second', 'second', 'second',\n",
       "       'second', 'second', 'second', 'second', 'second', 'second',\n",
       "       'second', 'second', 'second', 'second', 'second', 'second',\n",
       "       'second', 'None', 'None', 'None', 'None', 'None', 'None', 'None',\n",
       "       'None', 'None', 'None', 'None', 'None', 'None', 'None', 'None',\n",
       "       'None', 'None', 'None', 'None', 'None', 'None', 'None', 'None',\n",
       "       'None', 'None', 'None', 'None', 'None', 'None', 'None'],\n",
       "      dtype='<U6')"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.vectorize(map_dict.get)(df['kind'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "dict에 지정하지 않은 값들은 자신의 값 그대로 활용 가능"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['first', 'first', 'first', 'first', 'first', 'first', 'first',\n",
       "       'first', 'first', 'first', 'first', 'first', 'first', 'first',\n",
       "       'first', 'first', 'first', 'first', 'first', 'first', 'first',\n",
       "       'first', 'first', 'first', 'first', 'first', 'first', 'first',\n",
       "       'first', 'first', 'second', 'second', 'second', 'second', 'second',\n",
       "       'second', 'second', 'second', 'second', 'second', 'second',\n",
       "       'second', 'second', 'second', 'second', 'second', 'second',\n",
       "       'second', 'second', 'second', 'second', 'second', 'second',\n",
       "       'second', 'second', 'second', 'second', 'second', 'second',\n",
       "       'second', 'running', 'running', 'running', 'running', 'running',\n",
       "       'running', 'running', 'running', 'running', 'running', 'running',\n",
       "       'running', 'running', 'running', 'running', 'running', 'running',\n",
       "       'running', 'running', 'running', 'running', 'running', 'running',\n",
       "       'running', 'running', 'running', 'running', 'running', 'running',\n",
       "       'running'], dtype='<U7')"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.vectorize(map_dict.get)(df['kind'], df['kind'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "참고"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "611 µs ± 5.63 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)\n"
     ]
    }
   ],
   "source": [
    "%timeit df['kind'].map(map_dict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "165 µs ± 550 ns per loop (mean ± std. dev. of 7 runs, 10000 loops each)\n"
     ]
    }
   ],
   "source": [
    "%timeit np.vectorize(map_dict.get)(df['kind'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 데이터 조건에 따른 mapping"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### np.where 활용"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "np.where( 조건식, True 결과, False 결과 )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>diet</th>\n",
       "      <th>pulse</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>low fat</td>\n",
       "      <td>85</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>low fat</td>\n",
       "      <td>85</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>low fat</td>\n",
       "      <td>88</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2</td>\n",
       "      <td>low fat</td>\n",
       "      <td>90</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2</td>\n",
       "      <td>low fat</td>\n",
       "      <td>92</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   id     diet  pulse\n",
       "0   1  low fat     85\n",
       "1   1  low fat     85\n",
       "2   1  low fat     88\n",
       "3   2  low fat     90\n",
       "4   2  low fat     92"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = sns.load_dataset('exercise').iloc[:,[1,2,3]]\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 다중 조건에는 np.select 혹은 pd.cut or np.digitize 활용"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### pd.cut"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['result1'] = pd.cut(df['pulse'],bins=[-np.inf,70,90,np.inf], labels=['Bad','SoSo','Good'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "-np.inf < Bad <=70\n",
    "\n",
    "70 < Soso <= 90\n",
    "\n",
    "90 < Good < np.inf"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### np.digitize"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "right=True / False 조건을 pd.cut과 맞춰야 한다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "result = np.array(['Bad','SoSo','Good'])\n",
    "bins = np.digitize(df['pulse'].values, bins=[70, 90], right=True) # right 조건 체크"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['result2'] =  bins.choose(result)\n",
    "# bins.choose(result) == result[bins]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>diet</th>\n",
       "      <th>pulse</th>\n",
       "      <th>result2</th>\n",
       "      <th>result1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>low fat</td>\n",
       "      <td>85</td>\n",
       "      <td>SoSo</td>\n",
       "      <td>SoSo</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>low fat</td>\n",
       "      <td>85</td>\n",
       "      <td>SoSo</td>\n",
       "      <td>SoSo</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>low fat</td>\n",
       "      <td>88</td>\n",
       "      <td>SoSo</td>\n",
       "      <td>SoSo</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87</th>\n",
       "      <td>30</td>\n",
       "      <td>no fat</td>\n",
       "      <td>99</td>\n",
       "      <td>Good</td>\n",
       "      <td>Good</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>88</th>\n",
       "      <td>30</td>\n",
       "      <td>no fat</td>\n",
       "      <td>111</td>\n",
       "      <td>Good</td>\n",
       "      <td>Good</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>89</th>\n",
       "      <td>30</td>\n",
       "      <td>no fat</td>\n",
       "      <td>150</td>\n",
       "      <td>Good</td>\n",
       "      <td>Good</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>90 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    id     diet  pulse result2 result1\n",
       "0    1  low fat     85    SoSo    SoSo\n",
       "1    1  low fat     85    SoSo    SoSo\n",
       "2    1  low fat     88    SoSo    SoSo\n",
       "..  ..      ...    ...     ...     ...\n",
       "87  30   no fat     99    Good    Good\n",
       "88  30   no fat    111    Good    Good\n",
       "89  30   no fat    150    Good    Good\n",
       "\n",
       "[90 rows x 5 columns]"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### np.select"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "조건식 결과가 중복되지 않게 조절해야 함"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "condlist = [df['pulse'].values<70, df['pulse'].values>90] # 조건식 형태로 입력\n",
    "choicelist = ['Bad', 'Good']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['result'] = np.select(condlist, choicelist, default='SoSo')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>diet</th>\n",
       "      <th>pulse</th>\n",
       "      <th>result</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>low fat</td>\n",
       "      <td>85</td>\n",
       "      <td>SoSo</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>low fat</td>\n",
       "      <td>85</td>\n",
       "      <td>SoSo</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>low fat</td>\n",
       "      <td>88</td>\n",
       "      <td>SoSo</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87</th>\n",
       "      <td>30</td>\n",
       "      <td>no fat</td>\n",
       "      <td>99</td>\n",
       "      <td>Good</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>88</th>\n",
       "      <td>30</td>\n",
       "      <td>no fat</td>\n",
       "      <td>111</td>\n",
       "      <td>Good</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>89</th>\n",
       "      <td>30</td>\n",
       "      <td>no fat</td>\n",
       "      <td>150</td>\n",
       "      <td>Good</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>90 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    id     diet  pulse result\n",
       "0    1  low fat     85   SoSo\n",
       "1    1  low fat     85   SoSo\n",
       "2    1  low fat     88   SoSo\n",
       "..  ..      ...    ...    ...\n",
       "87  30   no fat     99   Good\n",
       "88  30   no fat    111   Good\n",
       "89  30   no fat    150   Good\n",
       "\n",
       "[90 rows x 4 columns]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}