doc/recurrent/keras_boston-LSTM.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import os, numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "os.environ[\"KERAS_BACKEND\"] = \"tensorflow\"\n",
    "\n",
    "import keras \n",
    "from sklearn import datasets\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "keras.backend.set_floatx('float64')\n",
    "\n",
    "from keras.utils import np_utils\n",
    "from keras.constraints import max_norm\n",
    "\n",
    "\n",
    "def set_reproducible():\n",
    "    import tensorflow as tf\n",
    "    import random as rn\n",
    "    import os\n",
    "    os.environ['PYTHONHASHSEED'] = '1960'\n",
    "    rn.seed(1960)\n",
    "    np.random.seed(1960)\n",
    "    session_conf = tf.ConfigProto(intra_op_parallelism_threads=1, inter_op_parallelism_threads=1)\n",
    "\n",
    "    from keras import backend as K\n",
    "    tf.set_random_seed(1960)\n",
    "\n",
    "    sess = tf.Session(graph=tf.get_default_graph(), config=session_conf)\n",
    "    # K.set_session(sess)\n",
    "\n",
    "set_reproducible()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Build a Keras Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "NC = 13\n",
    "\n",
    "def create_model():\n",
    "    from keras.models import Sequential\n",
    "    from keras.layers import Dense, Activation, LSTM\n",
    "    from keras.constraints import max_norm\n",
    "\n",
    "    model = Sequential()\n",
    "    model.add(LSTM(1, input_shape=(1, NC), batch_size=1 , \n",
    "                   activation='relu', recurrent_activation = 'hard_sigmoid', \n",
    "                   stateful=True,\n",
    "                   kernel_constraint=max_norm(2.),\n",
    "                   recurrent_constraint=max_norm(2.),\n",
    "                   kernel_initializer = keras.initializers.Orthogonal(gain=0.1),\n",
    "                   recurrent_initializer = keras.initializers.Zeros()))\n",
    "    model.add(Dense(1, activation=None))\n",
    "    adam_optimizer = keras.optimizers.Adam(lr=0.01, clipnorm=1.)\n",
    "    model.compile(loss='mse', optimizer=adam_optimizer)\n",
    "    return model\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "boston  = datasets.load_boston()\n",
    "boston.data = (boston.data - boston.data.mean(axis=0)) / boston.data.std(axis=0)\n",
    "boston.target = (boston.target - boston.target.mean(axis=0)) / boston.target.std(axis=0)\n",
    "\n",
    "boston.data = boston.data[:,0:NC]\n",
    "# boston.target = boston.data[:,1]\n",
    "\n",
    "train_X, test_X, train_y, test_y = train_test_split(boston.data, boston.target, train_size=0.8, test_size=0.2, random_state=1960)\n",
    "train_X = train_X.reshape(train_X.shape[0] , 1, train_X.shape[1])\n",
    "test_X = test_X.reshape(test_X.shape[0] , 1, test_X.shape[1])\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# train_X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "#boston.target"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(404, 1, 13) (404,) float64 float64\n",
      "Epoch 1/12\n",
      " - 1s - loss: 0.4152\n",
      "Epoch 2/12\n",
      " - 1s - loss: 0.3779\n",
      "Epoch 3/12\n",
      " - 1s - loss: 0.5066\n",
      "Epoch 4/12\n",
      " - 1s - loss: 0.5169\n",
      "Epoch 5/12\n",
      " - 1s - loss: 0.4786\n",
      "Epoch 6/12\n",
      " - 1s - loss: 0.5284\n",
      "Epoch 7/12\n",
      " - 1s - loss: 0.5092\n",
      "Epoch 8/12\n",
      " - 1s - loss: 0.4634\n",
      "Epoch 9/12\n",
      " - 1s - loss: 0.5034\n",
      "Epoch 10/12\n",
      " - 1s - loss: 0.5527\n",
      "Epoch 11/12\n",
      " - 1s - loss: 0.5294\n",
      "Epoch 12/12\n",
      " - 1s - loss: 0.5406\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x7f1aab4259e8>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "from keras.wrappers.scikit_learn import KerasRegressor\n",
    "\n",
    "clf = KerasRegressor(build_fn=create_model, batch_size=1, epochs=12, verbose=2)\n",
    "\n",
    "print(train_X.shape , train_y.shape, train_X.dtype , train_y.dtype)\n",
    "clf.fit(train_X, train_y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(102, 1, 13)\n",
      "[-0.13990413 -0.49634501 -0.3998435  -0.33324748  0.03652939 -0.49634501\n",
      " -0.49634501 -0.48721469 -0.49634501 -0.49634501 -0.49634501 -0.49634501\n",
      "  0.14033079 -0.49634501  1.52429822 -0.49634501  1.68447068 -0.49634501\n",
      " -0.49634501  1.15787203 -0.49634501  2.68152703 -0.49634501  0.68400461\n",
      " -0.49634501 -0.27459495 -0.49634501 -0.49634501  0.21643075 -0.49634501\n",
      "  0.56882751 -0.49634501 -0.49634501  1.53645988 -0.49634501  0.16216632\n",
      " -0.48807118 -0.49634501 -0.49623185  1.10702298 -0.49634501 -0.46756264\n",
      " -0.19462857 -0.49634501 -0.37754002 -0.3384696   0.00398761 -0.49634501\n",
      " -0.49634501 -0.49634501 -0.49560747 -0.49628708 -0.49634501  1.917483\n",
      " -0.49634501 -0.49634501  1.79298506 -0.49634501 -0.27103459 -0.47007884\n",
      " -0.45164406  1.73372738 -0.49634501  0.07662162 -0.49634501  1.65367332\n",
      " -0.49634501 -0.49634501 -0.49634501  0.58142615 -0.49634501  2.22291877\n",
      " -0.49634501 -0.240752   -0.07048631 -0.49634501  1.7257234  -0.49634501\n",
      " -0.49634501  1.26892235 -0.49634501  0.58954161 -0.49634501 -0.49634501\n",
      " -0.49634501  0.1214197  -0.49634501  1.89920441 -0.49634501  1.26233783\n",
      " -0.49634501  1.30275456 -0.49634501 -0.49634501  0.1447707  -0.49634501\n",
      "  1.85790451 -0.49634501  0.33802614 -0.49634501  0.73863981 -0.49634501]\n"
     ]
    }
   ],
   "source": [
    "print(test_X.shape)\n",
    "preds = clf.predict(test_X)\n",
    "clf.model.reset_states()\n",
    "print(preds)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "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>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>10</th>\n",
       "      <th>11</th>\n",
       "      <th>12</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>5.060000e+02</td>\n",
       "      <td>5.060000e+02</td>\n",
       "      <td>5.060000e+02</td>\n",
       "      <td>5.060000e+02</td>\n",
       "      <td>5.060000e+02</td>\n",
       "      <td>5.060000e+02</td>\n",
       "      <td>5.060000e+02</td>\n",
       "      <td>5.060000e+02</td>\n",
       "      <td>5.060000e+02</td>\n",
       "      <td>5.060000e+02</td>\n",
       "      <td>5.060000e+02</td>\n",
       "      <td>5.060000e+02</td>\n",
       "      <td>5.060000e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>-5.616939e-17</td>\n",
       "      <td>6.319056e-17</td>\n",
       "      <td>-3.145486e-15</td>\n",
       "      <td>-2.106352e-17</td>\n",
       "      <td>2.752300e-15</td>\n",
       "      <td>-1.150770e-14</td>\n",
       "      <td>-1.137430e-15</td>\n",
       "      <td>7.582867e-16</td>\n",
       "      <td>5.616939e-17</td>\n",
       "      <td>5.616939e-17</td>\n",
       "      <td>-1.022283e-14</td>\n",
       "      <td>8.593916e-15</td>\n",
       "      <td>-5.897786e-16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>1.000990e+00</td>\n",
       "      <td>1.000990e+00</td>\n",
       "      <td>1.000990e+00</td>\n",
       "      <td>1.000990e+00</td>\n",
       "      <td>1.000990e+00</td>\n",
       "      <td>1.000990e+00</td>\n",
       "      <td>1.000990e+00</td>\n",
       "      <td>1.000990e+00</td>\n",
       "      <td>1.000990e+00</td>\n",
       "      <td>1.000990e+00</td>\n",
       "      <td>1.000990e+00</td>\n",
       "      <td>1.000990e+00</td>\n",
       "      <td>1.000990e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>-4.197819e-01</td>\n",
       "      <td>-4.877224e-01</td>\n",
       "      <td>-1.557842e+00</td>\n",
       "      <td>-2.725986e-01</td>\n",
       "      <td>-1.465882e+00</td>\n",
       "      <td>-3.880249e+00</td>\n",
       "      <td>-2.335437e+00</td>\n",
       "      <td>-1.267069e+00</td>\n",
       "      <td>-9.828429e-01</td>\n",
       "      <td>-1.313990e+00</td>\n",
       "      <td>-2.707379e+00</td>\n",
       "      <td>-3.907193e+00</td>\n",
       "      <td>-1.531127e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>-4.109696e-01</td>\n",
       "      <td>-4.877224e-01</td>\n",
       "      <td>-8.676906e-01</td>\n",
       "      <td>-2.725986e-01</td>\n",
       "      <td>-9.130288e-01</td>\n",
       "      <td>-5.686303e-01</td>\n",
       "      <td>-8.374480e-01</td>\n",
       "      <td>-8.056878e-01</td>\n",
       "      <td>-6.379618e-01</td>\n",
       "      <td>-7.675760e-01</td>\n",
       "      <td>-4.880391e-01</td>\n",
       "      <td>2.050715e-01</td>\n",
       "      <td>-7.994200e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>-3.906665e-01</td>\n",
       "      <td>-4.877224e-01</td>\n",
       "      <td>-2.110985e-01</td>\n",
       "      <td>-2.725986e-01</td>\n",
       "      <td>-1.442174e-01</td>\n",
       "      <td>-1.084655e-01</td>\n",
       "      <td>3.173816e-01</td>\n",
       "      <td>-2.793234e-01</td>\n",
       "      <td>-5.230014e-01</td>\n",
       "      <td>-4.646726e-01</td>\n",
       "      <td>2.748590e-01</td>\n",
       "      <td>3.811865e-01</td>\n",
       "      <td>-1.812536e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>7.396560e-03</td>\n",
       "      <td>4.877224e-02</td>\n",
       "      <td>1.015999e+00</td>\n",
       "      <td>-2.725986e-01</td>\n",
       "      <td>5.986790e-01</td>\n",
       "      <td>4.827678e-01</td>\n",
       "      <td>9.067981e-01</td>\n",
       "      <td>6.623709e-01</td>\n",
       "      <td>1.661245e+00</td>\n",
       "      <td>1.530926e+00</td>\n",
       "      <td>8.065758e-01</td>\n",
       "      <td>4.336510e-01</td>\n",
       "      <td>6.030188e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>9.933931e+00</td>\n",
       "      <td>3.804234e+00</td>\n",
       "      <td>2.422565e+00</td>\n",
       "      <td>3.668398e+00</td>\n",
       "      <td>2.732346e+00</td>\n",
       "      <td>3.555044e+00</td>\n",
       "      <td>1.117494e+00</td>\n",
       "      <td>3.960518e+00</td>\n",
       "      <td>1.661245e+00</td>\n",
       "      <td>1.798194e+00</td>\n",
       "      <td>1.638828e+00</td>\n",
       "      <td>4.410519e-01</td>\n",
       "      <td>3.548771e+00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 0             1             2             3             4   \\\n",
       "count  5.060000e+02  5.060000e+02  5.060000e+02  5.060000e+02  5.060000e+02   \n",
       "mean  -5.616939e-17  6.319056e-17 -3.145486e-15 -2.106352e-17  2.752300e-15   \n",
       "std    1.000990e+00  1.000990e+00  1.000990e+00  1.000990e+00  1.000990e+00   \n",
       "min   -4.197819e-01 -4.877224e-01 -1.557842e+00 -2.725986e-01 -1.465882e+00   \n",
       "25%   -4.109696e-01 -4.877224e-01 -8.676906e-01 -2.725986e-01 -9.130288e-01   \n",
       "50%   -3.906665e-01 -4.877224e-01 -2.110985e-01 -2.725986e-01 -1.442174e-01   \n",
       "75%    7.396560e-03  4.877224e-02  1.015999e+00 -2.725986e-01  5.986790e-01   \n",
       "max    9.933931e+00  3.804234e+00  2.422565e+00  3.668398e+00  2.732346e+00   \n",
       "\n",
       "                 5             6             7             8             9   \\\n",
       "count  5.060000e+02  5.060000e+02  5.060000e+02  5.060000e+02  5.060000e+02   \n",
       "mean  -1.150770e-14 -1.137430e-15  7.582867e-16  5.616939e-17  5.616939e-17   \n",
       "std    1.000990e+00  1.000990e+00  1.000990e+00  1.000990e+00  1.000990e+00   \n",
       "min   -3.880249e+00 -2.335437e+00 -1.267069e+00 -9.828429e-01 -1.313990e+00   \n",
       "25%   -5.686303e-01 -8.374480e-01 -8.056878e-01 -6.379618e-01 -7.675760e-01   \n",
       "50%   -1.084655e-01  3.173816e-01 -2.793234e-01 -5.230014e-01 -4.646726e-01   \n",
       "75%    4.827678e-01  9.067981e-01  6.623709e-01  1.661245e+00  1.530926e+00   \n",
       "max    3.555044e+00  1.117494e+00  3.960518e+00  1.661245e+00  1.798194e+00   \n",
       "\n",
       "                 10            11            12  \n",
       "count  5.060000e+02  5.060000e+02  5.060000e+02  \n",
       "mean  -1.022283e-14  8.593916e-15 -5.897786e-16  \n",
       "std    1.000990e+00  1.000990e+00  1.000990e+00  \n",
       "min   -2.707379e+00 -3.907193e+00 -1.531127e+00  \n",
       "25%   -4.880391e-01  2.050715e-01 -7.994200e-01  \n",
       "50%    2.748590e-01  3.811865e-01 -1.812536e-01  \n",
       "75%    8.065758e-01  4.336510e-01  6.030188e-01  \n",
       "max    1.638828e+00  4.410519e-01  3.548771e+00  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame(boston.data).describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Generate SQL Code from the Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "import json, requests, base64, dill as pickle, sys\n",
    "\n",
    "\n",
    "\n",
    "sys.setrecursionlimit(200000)\n",
    "pickle.settings['recurse'] = False\n",
    "\n",
    "def test_ws_sql_gen(pickle_data):\n",
    "    WS_URL=\"https://sklearn2sql.herokuapp.com/model\"\n",
    "    b64_data = base64.b64encode(pickle_data).decode('utf-8')\n",
    "    data={\"Name\":\"model1\", \"PickleData\":b64_data , \"SQLDialect\":\"postgresql\"}\n",
    "    r = requests.post(WS_URL, json=data)\n",
    "    # print(r.__dict__)\n",
    "    content = r.json()\n",
    "    # print(content)\n",
    "    lSQL = content[\"model\"][\"SQLGenrationResult\"][0][\"SQL\"]\n",
    "    return lSQL;\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "pickle_data = pickle.dumps(clf)\n",
    "lSQL = test_ws_sql_gen(pickle_data)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WITH RECURSIVE keras_input AS \n",
      "(SELECT \"ADS\".\"KEY\" AS \"KEY\", \"ADS\".\"Feature_0\" AS \"Feature_0\", \"ADS\".\"Feature_1\" AS \"Feature_1\", \"ADS\".\"Feature_2\" AS \"Feature_2\", \"ADS\".\"Feature_3\" AS \"Feature_3\", \"ADS\".\"Feature_4\" AS \"Feature_4\", \"ADS\".\"Feature_5\" AS \"Feature_5\", \"ADS\".\"Feature_6\" AS \"Feature_6\", \"ADS\".\"Feature_7\" AS \"Feature_7\", \"ADS\".\"Feature_8\" AS \"Feature_8\", \"ADS\".\"Feature_9\" AS \"Feature_9\", \"ADS\".\"Feature_10\" AS \"Feature_10\", \"ADS\".\"Feature_11\" AS \"Feature_11\", \"ADS\".\"Feature_12\" AS \"Feature_12\" \n",
      "FROM \"INPUT_DATA\" AS \"ADS\"), \n",
      "keras_input_1 AS \n",
      "(SELECT keras_input.\"KEY\" AS \"KEY\", keras_input.\"Feature_0\" AS \"Feature_0\", keras_input.\"Feature_1\" AS \"Feature_1\", keras_input.\"Feature_2\" AS \"Feature_2\", keras_input.\"Feature_3\" AS \"Feature_3\", keras_input.\"Feature_4\" AS \"Feature_4\", keras_input.\"Feature_5\" AS \"Feature_5\", keras_input.\"Feature_6\" AS \"Feature_6\", keras_input.\"Feature_7\" AS \"Feature_7\", keras_input.\"Feature_8\" AS \"Feature_8\", keras_input.\"Feature_9\" AS \"Feature_9\", keras_input.\"Feature_10\" AS \"Feature_10\", keras_input.\"Feature_11\" AS \"Feature_11\", keras_input.\"Feature_12\" AS \"Feature_12\" \n",
      "FROM keras_input), \n",
      "keras_input_1_rn AS \n",
      "(SELECT row_number() OVER (ORDER BY keras_input_1.\"KEY\" ASC) AS rn, keras_input_1.\"KEY\" AS \"KEY\", keras_input_1.\"Feature_0\" AS \"Feature_0\", keras_input_1.\"Feature_1\" AS \"Feature_1\", keras_input_1.\"Feature_2\" AS \"Feature_2\", keras_input_1.\"Feature_3\" AS \"Feature_3\", keras_input_1.\"Feature_4\" AS \"Feature_4\", keras_input_1.\"Feature_5\" AS \"Feature_5\", keras_input_1.\"Feature_6\" AS \"Feature_6\", keras_input_1.\"Feature_7\" AS \"Feature_7\", keras_input_1.\"Feature_8\" AS \"Feature_8\", keras_input_1.\"Feature_9\" AS \"Feature_9\", keras_input_1.\"Feature_10\" AS \"Feature_10\", keras_input_1.\"Feature_11\" AS \"Feature_11\", keras_input_1.\"Feature_12\" AS \"Feature_12\" \n",
      "FROM keras_input_1), \n",
      "\"lstm_input_layer.kernel_bias\" AS \n",
      "(SELECT keras_input_1_rn.rn AS rn, keras_input_1_rn.\"KEY\" AS \"KEY\", keras_input_1_rn.\"Feature_0\" AS \"Feature_0\", keras_input_1_rn.\"Feature_1\" AS \"Feature_1\", keras_input_1_rn.\"Feature_2\" AS \"Feature_2\", keras_input_1_rn.\"Feature_3\" AS \"Feature_3\", keras_input_1_rn.\"Feature_4\" AS \"Feature_4\", keras_input_1_rn.\"Feature_5\" AS \"Feature_5\", keras_input_1_rn.\"Feature_6\" AS \"Feature_6\", keras_input_1_rn.\"Feature_7\" AS \"Feature_7\", keras_input_1_rn.\"Feature_8\" AS \"Feature_8\", keras_input_1_rn.\"Feature_9\" AS \"Feature_9\", keras_input_1_rn.\"Feature_10\" AS \"Feature_10\", keras_input_1_rn.\"Feature_11\" AS \"Feature_11\", keras_input_1_rn.\"Feature_12\" AS \"Feature_12\", 0.6475004717009399 + -0.2158820747194295 * keras_input_1_rn.\"Feature_0\" + 0.2238372261379932 * keras_input_1_rn.\"Feature_1\" + -0.25320975742283414 * keras_input_1_rn.\"Feature_2\" + 0.1672933744145996 * keras_input_1_rn.\"Feature_3\" + 0.2149260299246631 * keras_input_1_rn.\"Feature_4\" + 1.7351169684456662 * keras_input_1_rn.\"Feature_5\" + -0.2306316399798647 * keras_input_1_rn.\"Feature_6\" + -0.24047349548791963 * keras_input_1_rn.\"Feature_7\" + 0.2746322335390276 * keras_input_1_rn.\"Feature_8\" + -0.2895983740641028 * keras_input_1_rn.\"Feature_9\" + -0.516727711354856 * keras_input_1_rn.\"Feature_10\" + 0.13975814827769037 * keras_input_1_rn.\"Feature_11\" + -0.3928755012620301 * keras_input_1_rn.\"Feature_12\" AS dot_prod_i_1, -0.6102993476229848 + 0.7119030392498868 * keras_input_1_rn.\"Feature_0\" + -0.17416868418734888 * keras_input_1_rn.\"Feature_1\" + 0.7674353209278433 * keras_input_1_rn.\"Feature_2\" + -0.1521662813721242 * keras_input_1_rn.\"Feature_3\" + 0.6882096017471009 * keras_input_1_rn.\"Feature_4\" + 0.324764228648504 * keras_input_1_rn.\"Feature_5\" + -0.18652324183746016 * keras_input_1_rn.\"Feature_6\" + -0.36877495814665934 * keras_input_1_rn.\"Feature_7\" + 0.963852041076395 * keras_input_1_rn.\"Feature_8\" + 0.8510674564392534 * keras_input_1_rn.\"Feature_9\" + -0.06057484767029185 * keras_input_1_rn.\"Feature_10\" + -0.6475048523072433 * keras_input_1_rn.\"Feature_11\" + 0.15684124173575603 * keras_input_1_rn.\"Feature_12\" AS dot_prod_f_1, 1.5113562822561797 + -0.6360621560642756 * keras_input_1_rn.\"Feature_0\" + 0.07021960854110393 * keras_input_1_rn.\"Feature_1\" + -0.10648696146467895 * keras_input_1_rn.\"Feature_2\" + -0.06781430679956336 * keras_input_1_rn.\"Feature_3\" + -0.0234451999086851 * keras_input_1_rn.\"Feature_4\" + 1.3899058447142916 * keras_input_1_rn.\"Feature_5\" + -0.4178301029832567 * keras_input_1_rn.\"Feature_6\" + -0.2186342554333779 * keras_input_1_rn.\"Feature_7\" + -0.21527101228808798 * keras_input_1_rn.\"Feature_8\" + -0.46100394131537065 * keras_input_1_rn.\"Feature_9\" + -0.6706540874222692 * keras_input_1_rn.\"Feature_10\" + 0.39262699144907254 * keras_input_1_rn.\"Feature_11\" + -0.7463148651176916 * keras_input_1_rn.\"Feature_12\" AS dot_prod_c_1, 1.2813267317694246 + -0.664201821629264 * keras_input_1_rn.\"Feature_0\" + 0.03501540733563404 * keras_input_1_rn.\"Feature_1\" + -0.28926547768571875 * keras_input_1_rn.\"Feature_2\" + 0.3022299027378959 * keras_input_1_rn.\"Feature_3\" + -0.487447801195879 * keras_input_1_rn.\"Feature_4\" + 0.6748688764019186 * keras_input_1_rn.\"Feature_5\" + -0.5383467959964188 * keras_input_1_rn.\"Feature_6\" + -0.004024643047857606 * keras_input_1_rn.\"Feature_7\" + -0.1012163305970956 * keras_input_1_rn.\"Feature_8\" + -0.37218081783599916 * keras_input_1_rn.\"Feature_9\" + -0.6526910062496021 * keras_input_1_rn.\"Feature_10\" + 0.588824423046897 * keras_input_1_rn.\"Feature_11\" + -1.2158217400116675 * keras_input_1_rn.\"Feature_12\" AS dot_prod_o_1 \n",
      "FROM keras_input_1_rn), \n",
      "rnn_cte_lstm_1(rn_1, \"KEY\", \"PreviousState_1\", \"PreviousState_2\", \"State_1\", \"State_2\") AS \n",
      "(SELECT \"lstm_input_layer.kernel_bias\".rn AS rn_1, \"lstm_input_layer.kernel_bias\".\"KEY\" AS \"KEY\", CAST(0.0 AS FLOAT) AS \"PreviousState_1\", CAST(0.0 AS FLOAT) AS \"PreviousState_2\", greatest(0.0, least(1.0, 0.2 * (\"lstm_input_layer.kernel_bias\".dot_prod_o_1 + 0.0) + 0.5)) * greatest(greatest(0.0, least(1.0, 0.2 * (\"lstm_input_layer.kernel_bias\".dot_prod_i_1 + 0.0) + 0.5)) * greatest(\"lstm_input_layer.kernel_bias\".dot_prod_c_1 + 0.0, 0) + greatest(0.0, least(1.0, 0.2 * (\"lstm_input_layer.kernel_bias\".dot_prod_f_1 + 0.0) + 0.5)) * 0.0, 0) AS \"State_1\", greatest(0.0, least(1.0, 0.2 * (\"lstm_input_layer.kernel_bias\".dot_prod_i_1 + 0.0) + 0.5)) * greatest(\"lstm_input_layer.kernel_bias\".dot_prod_c_1 + 0.0, 0) + greatest(0.0, least(1.0, 0.2 * (\"lstm_input_layer.kernel_bias\".dot_prod_f_1 + 0.0) + 0.5)) * 0.0 AS \"State_2\" \n",
      "FROM \"lstm_input_layer.kernel_bias\" \n",
      "WHERE \"lstm_input_layer.kernel_bias\".rn = 1 UNION ALL SELECT \"lstm_input_layer.kernel_bias\".rn AS rn, \"lstm_input_layer.kernel_bias\".\"KEY\" AS \"KEY\", rnn_cte_lstm_1.\"State_1\" AS \"PreviousState_1\", rnn_cte_lstm_1.\"State_2\" AS \"PreviousState_2\", greatest(0.0, least(1.0, 0.2 * (\"lstm_input_layer.kernel_bias\".dot_prod_o_1 + -1.494396033820323 * rnn_cte_lstm_1.\"State_1\") + 0.5)) * greatest(greatest(0.0, least(1.0, 0.2 * (\"lstm_input_layer.kernel_bias\".dot_prod_i_1 + -1.9969204331168193 * rnn_cte_lstm_1.\"State_1\") + 0.5)) * greatest(\"lstm_input_layer.kernel_bias\".dot_prod_c_1 + -1.960457767080738 * rnn_cte_lstm_1.\"State_1\", 0) + greatest(0.0, least(1.0, 0.2 * (\"lstm_input_layer.kernel_bias\".dot_prod_f_1 + -1.9400724565523353 * rnn_cte_lstm_1.\"State_1\") + 0.5)) * rnn_cte_lstm_1.\"State_2\", 0) AS \"State_1\", greatest(0.0, least(1.0, 0.2 * (\"lstm_input_layer.kernel_bias\".dot_prod_i_1 + -1.9969204331168193 * rnn_cte_lstm_1.\"State_1\") + 0.5)) * greatest(\"lstm_input_layer.kernel_bias\".dot_prod_c_1 + -1.960457767080738 * rnn_cte_lstm_1.\"State_1\", 0) + greatest(0.0, least(1.0, 0.2 * (\"lstm_input_layer.kernel_bias\".dot_prod_f_1 + -1.9400724565523353 * rnn_cte_lstm_1.\"State_1\") + 0.5)) * rnn_cte_lstm_1.\"State_2\" AS \"State_2\" \n",
      "FROM \"lstm_input_layer.kernel_bias\", rnn_cte_lstm_1 \n",
      "WHERE \"lstm_input_layer.kernel_bias\".rn = rnn_cte_lstm_1.rn_1 + 1), \n",
      "lstm_1 AS \n",
      "(SELECT rnn_cte_lstm_1.\"KEY\" AS \"KEY\", CAST(rnn_cte_lstm_1.\"State_1\" AS FLOAT) AS output_1 \n",
      "FROM rnn_cte_lstm_1), \n",
      "lstm_1_1 AS \n",
      "(SELECT lstm_1.\"KEY\" AS \"KEY\", lstm_1.output_1 AS output_1 \n",
      "FROM lstm_1), \n",
      "layer_dense_1 AS \n",
      "(SELECT lstm_1_1.\"KEY\" AS \"KEY\", -0.4963450083727642 + 0.37008790284834636 * lstm_1_1.output_1 AS output_1 \n",
      "FROM lstm_1_1)\n",
      " SELECT layer_dense_1.\"KEY\" AS \"KEY\", layer_dense_1.output_1 AS \"Estimator\" \n",
      "FROM layer_dense_1\n"
     ]
    }
   ],
   "source": [
    "print(lSQL)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Execute the SQL Code"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/antoine/.local/lib/python3.7/site-packages/psycopg2/__init__.py:144: UserWarning: The psycopg2 wheel package will be renamed from release 2.8; in order to keep installing from binary please use \"pip install psycopg2-binary\" instead. For details see: <http://initd.org/psycopg/docs/install.html#binary-install-from-pypi>.\n",
      "  \"\"\")\n"
     ]
    }
   ],
   "source": [
    "# save the dataset in a database table\n",
    "\n",
    "\n",
    "import sqlalchemy as sa\n",
    "\n",
    "#engine = sa.create_engine('sqlite://' , echo=False)\n",
    "engine = sa.create_engine(\"postgresql://db:db@localhost/db?port=5432\", echo=False)\n",
    "conn = engine.connect()\n",
    "\n",
    "lTable = pd.DataFrame(boston.data);\n",
    "lTable.columns = ['Feature_' + str(x) for x in range(boston.data.shape[1]) ]\n",
    "lTable['TGT'] = boston.target\n",
    "lTable['KEY'] = range(boston.data.shape[0])\n",
    "lTable.to_sql(\"INPUT_DATA\" , conn,   if_exists='replace', index=False)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "sql_output = pd.read_sql(lSQL , conn);\n",
    "sql_output = sql_output.sort_values(by='KEY').reset_index(drop=True)\n",
    "conn.close()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "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>KEY</th>\n",
       "      <th>Estimator</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>230</th>\n",
       "      <td>230</td>\n",
       "      <td>-0.059490</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>112</th>\n",
       "      <td>112</td>\n",
       "      <td>-0.496345</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>125</th>\n",
       "      <td>125</td>\n",
       "      <td>-0.411700</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>9</td>\n",
       "      <td>-0.185947</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>213</th>\n",
       "      <td>213</td>\n",
       "      <td>-0.018767</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>109</th>\n",
       "      <td>109</td>\n",
       "      <td>-0.494333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>127</th>\n",
       "      <td>127</td>\n",
       "      <td>-0.493254</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>244</th>\n",
       "      <td>244</td>\n",
       "      <td>-0.496345</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>406</th>\n",
       "      <td>406</td>\n",
       "      <td>-0.496345</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>490</th>\n",
       "      <td>490</td>\n",
       "      <td>-0.496345</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>245</th>\n",
       "      <td>245</td>\n",
       "      <td>-0.496345</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>133</th>\n",
       "      <td>133</td>\n",
       "      <td>-0.483684</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     KEY  Estimator\n",
       "230  230  -0.059490\n",
       "112  112  -0.496345\n",
       "125  125  -0.411700\n",
       "9      9  -0.185947\n",
       "213  213  -0.018767\n",
       "109  109  -0.494333\n",
       "127  127  -0.493254\n",
       "244  244  -0.496345\n",
       "406  406  -0.496345\n",
       "490  490  -0.496345\n",
       "245  245  -0.496345\n",
       "133  133  -0.483684"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sql_output.sample(12, random_state=1960)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Keras Prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(506,)\n"
     ]
    },
    {
     "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>KEY</th>\n",
       "      <th>Estimator</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>230</th>\n",
       "      <td>230</td>\n",
       "      <td>-0.059490</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>112</th>\n",
       "      <td>112</td>\n",
       "      <td>-0.496345</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>125</th>\n",
       "      <td>125</td>\n",
       "      <td>-0.411700</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>9</td>\n",
       "      <td>-0.185947</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>213</th>\n",
       "      <td>213</td>\n",
       "      <td>-0.018767</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>109</th>\n",
       "      <td>109</td>\n",
       "      <td>-0.494333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>127</th>\n",
       "      <td>127</td>\n",
       "      <td>-0.493254</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>244</th>\n",
       "      <td>244</td>\n",
       "      <td>-0.496345</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>406</th>\n",
       "      <td>406</td>\n",
       "      <td>-0.496345</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>490</th>\n",
       "      <td>490</td>\n",
       "      <td>-0.496345</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>245</th>\n",
       "      <td>245</td>\n",
       "      <td>-0.496345</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>133</th>\n",
       "      <td>133</td>\n",
       "      <td>-0.483684</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     KEY  Estimator\n",
       "230  230  -0.059490\n",
       "112  112  -0.496345\n",
       "125  125  -0.411700\n",
       "9      9  -0.185947\n",
       "213  213  -0.018767\n",
       "109  109  -0.494333\n",
       "127  127  -0.493254\n",
       "244  244  -0.496345\n",
       "406  406  -0.496345\n",
       "490  490  -0.496345\n",
       "245  245  -0.496345\n",
       "133  133  -0.483684"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#clf.model.reset_states()\n",
    "skl_output = pd.DataFrame()\n",
    "skl_output_key = pd.DataFrame(list(range(boston.data.shape[0])), columns=['KEY']);\n",
    "data = boston.data.reshape(boston.data.shape[0] , 1, boston.data.shape[1])\n",
    "skl_out = clf.predict(data)\n",
    "print(skl_out.shape)\n",
    "skl_output_estimator = pd.DataFrame(skl_out, columns=['Estimator'])\n",
    "skl_output['KEY'] = skl_output_key['KEY']\n",
    "skl_output['Estimator'] = skl_output_estimator\n",
    "skl_output.sample(12, random_state=1960)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "format": "column"
   },
   "source": [
    "# Comparing the SQL and Keras Predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "sql_skl_join = skl_output.join(sql_output , how='left', on='KEY', lsuffix='_skl', rsuffix='_sql')\n",
    "sql_skl_join['Error'] = sql_skl_join.Estimator_sql - sql_skl_join.Estimator_skl"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "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>KEY_skl</th>\n",
       "      <th>Estimator_skl</th>\n",
       "      <th>KEY_sql</th>\n",
       "      <th>Estimator_sql</th>\n",
       "      <th>Error</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>1.363376</td>\n",
       "      <td>0</td>\n",
       "      <td>1.363376</td>\n",
       "      <td>-3.330669e-15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>-0.496345</td>\n",
       "      <td>1</td>\n",
       "      <td>-0.496345</td>\n",
       "      <td>1.665335e-16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>1.545318</td>\n",
       "      <td>2</td>\n",
       "      <td>1.545318</td>\n",
       "      <td>-3.330669e-15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>-0.496345</td>\n",
       "      <td>3</td>\n",
       "      <td>-0.496345</td>\n",
       "      <td>1.665335e-16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>1.405068</td>\n",
       "      <td>4</td>\n",
       "      <td>1.405068</td>\n",
       "      <td>-1.110223e-15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>5</td>\n",
       "      <td>-0.496345</td>\n",
       "      <td>5</td>\n",
       "      <td>-0.496345</td>\n",
       "      <td>1.665335e-16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>6</td>\n",
       "      <td>0.167765</td>\n",
       "      <td>6</td>\n",
       "      <td>0.167765</td>\n",
       "      <td>-5.551115e-17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>7</td>\n",
       "      <td>-0.496345</td>\n",
       "      <td>7</td>\n",
       "      <td>-0.496345</td>\n",
       "      <td>1.665335e-16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>8</td>\n",
       "      <td>-0.496345</td>\n",
       "      <td>8</td>\n",
       "      <td>-0.496345</td>\n",
       "      <td>1.665335e-16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>9</td>\n",
       "      <td>-0.185947</td>\n",
       "      <td>9</td>\n",
       "      <td>-0.185947</td>\n",
       "      <td>-2.775558e-17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>10</td>\n",
       "      <td>-0.464172</td>\n",
       "      <td>10</td>\n",
       "      <td>-0.464172</td>\n",
       "      <td>4.996004e-16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>11</td>\n",
       "      <td>-0.043747</td>\n",
       "      <td>11</td>\n",
       "      <td>-0.043747</td>\n",
       "      <td>-4.163336e-17</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    KEY_skl  Estimator_skl  KEY_sql  Estimator_sql         Error\n",
       "0         0       1.363376        0       1.363376 -3.330669e-15\n",
       "1         1      -0.496345        1      -0.496345  1.665335e-16\n",
       "2         2       1.545318        2       1.545318 -3.330669e-15\n",
       "3         3      -0.496345        3      -0.496345  1.665335e-16\n",
       "4         4       1.405068        4       1.405068 -1.110223e-15\n",
       "5         5      -0.496345        5      -0.496345  1.665335e-16\n",
       "6         6       0.167765        6       0.167765 -5.551115e-17\n",
       "7         7      -0.496345        7      -0.496345  1.665335e-16\n",
       "8         8      -0.496345        8      -0.496345  1.665335e-16\n",
       "9         9      -0.185947        9      -0.185947 -2.775558e-17\n",
       "10       10      -0.464172       10      -0.464172  4.996004e-16\n",
       "11       11      -0.043747       11      -0.043747 -4.163336e-17"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sql_skl_join.head(12)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "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>KEY_skl</th>\n",
       "      <th>Estimator_skl</th>\n",
       "      <th>KEY_sql</th>\n",
       "      <th>Estimator_sql</th>\n",
       "      <th>Error</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>506.000000</td>\n",
       "      <td>506.000000</td>\n",
       "      <td>506.000000</td>\n",
       "      <td>506.000000</td>\n",
       "      <td>5.060000e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>252.500000</td>\n",
       "      <td>-0.107125</td>\n",
       "      <td>252.500000</td>\n",
       "      <td>-0.107125</td>\n",
       "      <td>9.535837e-17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>146.213884</td>\n",
       "      <td>0.718815</td>\n",
       "      <td>146.213884</td>\n",
       "      <td>0.718815</td>\n",
       "      <td>9.630438e-16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>-0.496345</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-0.496345</td>\n",
       "      <td>-5.329071e-15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>126.250000</td>\n",
       "      <td>-0.496345</td>\n",
       "      <td>126.250000</td>\n",
       "      <td>-0.496345</td>\n",
       "      <td>-5.551115e-17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>252.500000</td>\n",
       "      <td>-0.490673</td>\n",
       "      <td>252.500000</td>\n",
       "      <td>-0.490673</td>\n",
       "      <td>1.665335e-16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>378.750000</td>\n",
       "      <td>-0.124858</td>\n",
       "      <td>378.750000</td>\n",
       "      <td>-0.124858</td>\n",
       "      <td>1.665335e-16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>505.000000</td>\n",
       "      <td>2.864752</td>\n",
       "      <td>505.000000</td>\n",
       "      <td>2.864752</td>\n",
       "      <td>4.884981e-15</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          KEY_skl  Estimator_skl     KEY_sql  Estimator_sql         Error\n",
       "count  506.000000     506.000000  506.000000     506.000000  5.060000e+02\n",
       "mean   252.500000      -0.107125  252.500000      -0.107125  9.535837e-17\n",
       "std    146.213884       0.718815  146.213884       0.718815  9.630438e-16\n",
       "min      0.000000      -0.496345    0.000000      -0.496345 -5.329071e-15\n",
       "25%    126.250000      -0.496345  126.250000      -0.496345 -5.551115e-17\n",
       "50%    252.500000      -0.490673  252.500000      -0.490673  1.665335e-16\n",
       "75%    378.750000      -0.124858  378.750000      -0.124858  1.665335e-16\n",
       "max    505.000000       2.864752  505.000000       2.864752  4.884981e-15"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sql_skl_join.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}