HDFSQuota.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from os import listdir\n",
    "from os.path import isfile\n",
    "import pandas as pd\n",
    "import matplotlib as mpl\n",
    "import matplotlib.pyplot as plt\n",
    "import brewer2mpl\n",
    "import matplotlib.font_manager\n",
    "%matplotlib inline\n",
    "# args ( name, type, number of colors)\n",
    "pd.set_option('display.mpl_style', 'default')\n",
    "bmap = brewer2mpl.get_map('Set2', 'qualitative', 7)\n",
    "mpl.rcParams['axes.color_cycle'] = bmap.mpl_colors\n",
    "mpl.rcParams['font.size'] = 14\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['hdfs-dfs-lsr-20151024.log']\n"
     ]
    }
   ],
   "source": [
    "# get hdfs log files\n",
    "hdfs_files = [f for f in listdir() if isfile(f) & f.startswith('hdfs') & f.endswith('.log')]   \n",
    "print(hdfs_files)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>permissions</th>\n",
       "      <th>replicas</th>\n",
       "      <th>user</th>\n",
       "      <th>group</th>\n",
       "      <th>size</th>\n",
       "      <th>mod_date</th>\n",
       "      <th>mod_time</th>\n",
       "      <th>fname</th>\n",
       "      <th>src</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>drwxrwxrwx</td>\n",
       "      <td>-</td>\n",
       "      <td>yarn</td>\n",
       "      <td>hadoop</td>\n",
       "      <td>0</td>\n",
       "      <td>2015-10-20</td>\n",
       "      <td>13:26</td>\n",
       "      <td>/app-logs</td>\n",
       "      <td>20151024</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>drwxrwx---</td>\n",
       "      <td>-</td>\n",
       "      <td>ambari-qa</td>\n",
       "      <td>hadoop</td>\n",
       "      <td>0</td>\n",
       "      <td>2015-10-20</td>\n",
       "      <td>13:26</td>\n",
       "      <td>/app-logs/ambari-qa</td>\n",
       "      <td>20151024</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>drwxrwx---</td>\n",
       "      <td>-</td>\n",
       "      <td>ambari-qa</td>\n",
       "      <td>hadoop</td>\n",
       "      <td>0</td>\n",
       "      <td>2015-10-21</td>\n",
       "      <td>00:47</td>\n",
       "      <td>/app-logs/ambari-qa/logs</td>\n",
       "      <td>20151024</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>drwxrwx---</td>\n",
       "      <td>-</td>\n",
       "      <td>ambari-qa</td>\n",
       "      <td>hadoop</td>\n",
       "      <td>0</td>\n",
       "      <td>2015-10-20</td>\n",
       "      <td>13:26</td>\n",
       "      <td>/app-logs/ambari-qa/logs/application_144537273...</td>\n",
       "      <td>20151024</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>-rw-r-----</td>\n",
       "      <td>2</td>\n",
       "      <td>ambari-qa</td>\n",
       "      <td>hadoop</td>\n",
       "      <td>384</td>\n",
       "      <td>2015-10-20</td>\n",
       "      <td>13:26</td>\n",
       "      <td>/app-logs/ambari-qa/logs/application_144537273...</td>\n",
       "      <td>20151024</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  permissions replicas       user   group  size    mod_date mod_time  \\\n",
       "0  drwxrwxrwx        -       yarn  hadoop     0  2015-10-20    13:26   \n",
       "1  drwxrwx---        -  ambari-qa  hadoop     0  2015-10-20    13:26   \n",
       "2  drwxrwx---        -  ambari-qa  hadoop     0  2015-10-21    00:47   \n",
       "3  drwxrwx---        -  ambari-qa  hadoop     0  2015-10-20    13:26   \n",
       "4  -rw-r-----        2  ambari-qa  hadoop   384  2015-10-20    13:26   \n",
       "\n",
       "                                               fname       src  \n",
       "0                                          /app-logs  20151024  \n",
       "1                                /app-logs/ambari-qa  20151024  \n",
       "2                           /app-logs/ambari-qa/logs  20151024  \n",
       "3  /app-logs/ambari-qa/logs/application_144537273...  20151024  \n",
       "4  /app-logs/ambari-qa/logs/application_144537273...  20151024  "
      ]
     },
     "execution_count": 87,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# only read in first log file for now (in the future read in more files to compare different snapshots)\n",
    "hdfs_files = ['hdfs-dfs-lsr-20151024.log']\n",
    "df = pd.DataFrame()\n",
    "for f in hdfs_files:\n",
    "    dft = pd.read_csv(f, delim_whitespace=True, names=['permissions', 'replicas', 'user', 'group', 'size', 'mod_date', 'mod_time', 'fname'])\n",
    "    dft['src'] = f.replace('hdfs-dfs-lsr-','').replace('.log','')\n",
    "    df = pd.concat([df, dft])\n",
    "\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "#df = pd.read_csv('hdfs-dfs-lsr-1.log', delim_whitespace=True, names=['permissions', 'replicas', 'user', 'group', 'size', 'mod_date', 'mod_time', 'fname'])\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>permissions</th>\n",
       "      <th>replicas</th>\n",
       "      <th>user</th>\n",
       "      <th>group</th>\n",
       "      <th>size</th>\n",
       "      <th>mod_date</th>\n",
       "      <th>mod_time</th>\n",
       "      <th>fname</th>\n",
       "      <th>src</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>drwxrwxrwx</td>\n",
       "      <td>-</td>\n",
       "      <td>yarn</td>\n",
       "      <td>hadoop</td>\n",
       "      <td>0</td>\n",
       "      <td>2015-10-20</td>\n",
       "      <td>13:26</td>\n",
       "      <td>/app-logs</td>\n",
       "      <td>20151024</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>drwxrwx---</td>\n",
       "      <td>-</td>\n",
       "      <td>ambari-qa</td>\n",
       "      <td>hadoop</td>\n",
       "      <td>0</td>\n",
       "      <td>2015-10-20</td>\n",
       "      <td>13:26</td>\n",
       "      <td>/app-logs/ambari-qa</td>\n",
       "      <td>20151024</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>drwxrwx---</td>\n",
       "      <td>-</td>\n",
       "      <td>ambari-qa</td>\n",
       "      <td>hadoop</td>\n",
       "      <td>0</td>\n",
       "      <td>2015-10-21</td>\n",
       "      <td>00:47</td>\n",
       "      <td>/app-logs/ambari-qa/logs</td>\n",
       "      <td>20151024</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>drwxrwx---</td>\n",
       "      <td>-</td>\n",
       "      <td>ambari-qa</td>\n",
       "      <td>hadoop</td>\n",
       "      <td>0</td>\n",
       "      <td>2015-10-20</td>\n",
       "      <td>13:26</td>\n",
       "      <td>/app-logs/ambari-qa/logs/application_144537273...</td>\n",
       "      <td>20151024</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>-rw-r-----</td>\n",
       "      <td>2</td>\n",
       "      <td>ambari-qa</td>\n",
       "      <td>hadoop</td>\n",
       "      <td>384</td>\n",
       "      <td>2015-10-20</td>\n",
       "      <td>13:26</td>\n",
       "      <td>/app-logs/ambari-qa/logs/application_144537273...</td>\n",
       "      <td>20151024</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  permissions replicas       user   group  size    mod_date mod_time  \\\n",
       "0  drwxrwxrwx        -       yarn  hadoop     0  2015-10-20    13:26   \n",
       "1  drwxrwx---        -  ambari-qa  hadoop     0  2015-10-20    13:26   \n",
       "2  drwxrwx---        -  ambari-qa  hadoop     0  2015-10-21    00:47   \n",
       "3  drwxrwx---        -  ambari-qa  hadoop     0  2015-10-20    13:26   \n",
       "4  -rw-r-----        2  ambari-qa  hadoop   384  2015-10-20    13:26   \n",
       "\n",
       "                                               fname       src  \n",
       "0                                          /app-logs  20151024  \n",
       "1                                /app-logs/ambari-qa  20151024  \n",
       "2                           /app-logs/ambari-qa/logs  20151024  \n",
       "3  /app-logs/ambari-qa/logs/application_144537273...  20151024  \n",
       "4  /app-logs/ambari-qa/logs/application_144537273...  20151024  "
      ]
     },
     "execution_count": 89,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>permissions</th>\n",
       "      <th>replicas</th>\n",
       "      <th>user</th>\n",
       "      <th>group</th>\n",
       "      <th>size</th>\n",
       "      <th>mod_date</th>\n",
       "      <th>mod_time</th>\n",
       "      <th>fname</th>\n",
       "      <th>src</th>\n",
       "      <th>is_dir</th>\n",
       "      <th>num_replicas</th>\n",
       "      <th>dt_mod</th>\n",
       "      <th>size_mb</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>drwxrwxrwx</td>\n",
       "      <td>-</td>\n",
       "      <td>yarn</td>\n",
       "      <td>hadoop</td>\n",
       "      <td>0</td>\n",
       "      <td>2015-10-20</td>\n",
       "      <td>13:26</td>\n",
       "      <td>/app-logs</td>\n",
       "      <td>20151024</td>\n",
       "      <td>True</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2015-10-20 13:26:00</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>drwxrwx---</td>\n",
       "      <td>-</td>\n",
       "      <td>ambari-qa</td>\n",
       "      <td>hadoop</td>\n",
       "      <td>0</td>\n",
       "      <td>2015-10-20</td>\n",
       "      <td>13:26</td>\n",
       "      <td>/app-logs/ambari-qa</td>\n",
       "      <td>20151024</td>\n",
       "      <td>True</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2015-10-20 13:26:00</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>drwxrwx---</td>\n",
       "      <td>-</td>\n",
       "      <td>ambari-qa</td>\n",
       "      <td>hadoop</td>\n",
       "      <td>0</td>\n",
       "      <td>2015-10-21</td>\n",
       "      <td>00:47</td>\n",
       "      <td>/app-logs/ambari-qa/logs</td>\n",
       "      <td>20151024</td>\n",
       "      <td>True</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2015-10-21 00:47:00</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>drwxrwx---</td>\n",
       "      <td>-</td>\n",
       "      <td>ambari-qa</td>\n",
       "      <td>hadoop</td>\n",
       "      <td>0</td>\n",
       "      <td>2015-10-20</td>\n",
       "      <td>13:26</td>\n",
       "      <td>/app-logs/ambari-qa/logs/application_144537273...</td>\n",
       "      <td>20151024</td>\n",
       "      <td>True</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2015-10-20 13:26:00</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>-rw-r-----</td>\n",
       "      <td>2</td>\n",
       "      <td>ambari-qa</td>\n",
       "      <td>hadoop</td>\n",
       "      <td>384</td>\n",
       "      <td>2015-10-20</td>\n",
       "      <td>13:26</td>\n",
       "      <td>/app-logs/ambari-qa/logs/application_144537273...</td>\n",
       "      <td>20151024</td>\n",
       "      <td>False</td>\n",
       "      <td>2</td>\n",
       "      <td>2015-10-20 13:26:00</td>\n",
       "      <td>0.000366</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  permissions replicas       user   group  size    mod_date mod_time  \\\n",
       "0  drwxrwxrwx        -       yarn  hadoop     0  2015-10-20    13:26   \n",
       "1  drwxrwx---        -  ambari-qa  hadoop     0  2015-10-20    13:26   \n",
       "2  drwxrwx---        -  ambari-qa  hadoop     0  2015-10-21    00:47   \n",
       "3  drwxrwx---        -  ambari-qa  hadoop     0  2015-10-20    13:26   \n",
       "4  -rw-r-----        2  ambari-qa  hadoop   384  2015-10-20    13:26   \n",
       "\n",
       "                                               fname       src is_dir  \\\n",
       "0                                          /app-logs  20151024   True   \n",
       "1                                /app-logs/ambari-qa  20151024   True   \n",
       "2                           /app-logs/ambari-qa/logs  20151024   True   \n",
       "3  /app-logs/ambari-qa/logs/application_144537273...  20151024   True   \n",
       "4  /app-logs/ambari-qa/logs/application_144537273...  20151024  False   \n",
       "\n",
       "   num_replicas              dt_mod   size_mb  \n",
       "0           NaN 2015-10-20 13:26:00  0.000000  \n",
       "1           NaN 2015-10-20 13:26:00  0.000000  \n",
       "2           NaN 2015-10-21 00:47:00  0.000000  \n",
       "3           NaN 2015-10-20 13:26:00  0.000000  \n",
       "4             2 2015-10-20 13:26:00  0.000366  "
      ]
     },
     "execution_count": 90,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['is_dir'] = df['permissions'].apply(lambda x: True if x[0] == 'd' else False)\n",
    "df['num_replicas'] = pd.to_numeric(df.replicas,'coerce')\n",
    "df['dt_mod'] = pd.to_datetime(df.mod_date + ' ' + df.mod_time)\n",
    "df['size_mb'] = df['size'].apply(lambda size: size/1024/1024)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1708"
      ]
     },
     "execution_count": 91,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# num files\n",
    "df[(df.is_dir==False)]['fname'].count()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "is_dir\n",
       "False    1708\n",
       "True     1029\n",
       "Name: fname, dtype: int64"
      ]
     },
     "execution_count": 92,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# num files and directories\n",
    "# True=Directory; False=File\n",
    "df.groupby(['is_dir'])['fname'].count()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "user       group   is_dir\n",
       "ambari-qa  hadoop  False      91\n",
       "                   True       39\n",
       "           hdfs    False      42\n",
       "                   True       25\n",
       "anonymous  hdfs    True        1\n",
       "hbase      hdfs    False     604\n",
       "                   True      827\n",
       "hcat       hdfs    True        1\n",
       "hdfs       hadoop  False      10\n",
       "           hdfs    False     139\n",
       "                   True       94\n",
       "hive       hdfs    True        7\n",
       "mapred     hadoop  True        9\n",
       "           hdfs    True        1\n",
       "oozie      hadoop  True        1\n",
       "           hdfs    False     822\n",
       "                   True       22\n",
       "spark      hdfs    True        1\n",
       "yarn       hadoop  True        1\n",
       "Name: fname, dtype: int64"
      ]
     },
     "execution_count": 93,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#num files/dirs per user:group\n",
    "df.groupby(['user','group','is_dir'])['fname'].count()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "user       group \n",
       "ambari-qa  hadoop      3.866881\n",
       "           hdfs       20.353203\n",
       "anonymous  hdfs        0.000000\n",
       "hbase      hdfs        0.388103\n",
       "hcat       hdfs        0.000000\n",
       "hdfs       hadoop    872.678347\n",
       "           hdfs        8.032172\n",
       "hive       hdfs        0.000000\n",
       "mapred     hadoop      0.000000\n",
       "           hdfs        0.000000\n",
       "oozie      hadoop      0.000000\n",
       "           hdfs      602.341820\n",
       "spark      hdfs        0.000000\n",
       "yarn       hadoop      0.000000\n",
       "Name: size_mb, dtype: float64"
      ]
     },
     "execution_count": 94,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# quota per user:group\n",
    "df.groupby(['user','group'])['size_mb'].sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "user       group \n",
       "ambari-qa  hadoop   2015-10-21 00:48:00\n",
       "           hdfs     2015-10-21 11:16:00\n",
       "anonymous  hdfs     2015-10-20 22:13:00\n",
       "hbase      hdfs     2015-10-24 05:43:00\n",
       "hcat       hdfs     2015-10-20 06:30:00\n",
       "hdfs       hadoop   2015-10-21 00:25:00\n",
       "           hdfs     2015-10-23 14:32:00\n",
       "hive       hdfs     2015-10-21 11:16:00\n",
       "mapred     hadoop   2015-10-21 00:47:00\n",
       "           hdfs     2015-10-20 06:29:00\n",
       "oozie      hadoop   2015-10-20 13:27:00\n",
       "           hdfs     2015-10-21 00:31:00\n",
       "spark      hdfs     2015-10-20 06:28:00\n",
       "yarn       hadoop   2015-10-20 13:26:00\n",
       "Name: dt_mod, dtype: datetime64[ns]"
      ]
     },
     "execution_count": 95,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# last seen (user:group)\n",
    "df.groupby(['user','group'])['dt_mod'].max()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1607   2015-10-23 14:32:00\n",
       "1595   2015-10-23 14:32:00\n",
       "1605   2015-10-23 14:32:00\n",
       "1606   2015-10-23 14:32:00\n",
       "1600   2015-10-22 14:32:00\n",
       "1604   2015-10-22 14:32:00\n",
       "1601   2015-10-21 15:30:00\n",
       "1602   2015-10-21 15:30:00\n",
       "1603   2015-10-21 15:30:00\n",
       "1599   2015-10-21 14:32:00\n",
       "85     2015-10-21 11:06:00\n",
       "84     2015-10-21 11:06:00\n",
       "1597   2015-10-21 01:01:00\n",
       "1608   2015-10-21 00:49:00\n",
       "1786   2015-10-21 00:44:00\n",
       "1711   2015-10-21 00:44:00\n",
       "1710   2015-10-21 00:44:00\n",
       "1709   2015-10-21 00:44:00\n",
       "1708   2015-10-21 00:44:00\n",
       "1707   2015-10-21 00:44:00\n",
       "1706   2015-10-21 00:44:00\n",
       "1783   2015-10-21 00:44:00\n",
       "1784   2015-10-21 00:44:00\n",
       "1785   2015-10-21 00:44:00\n",
       "1787   2015-10-21 00:44:00\n",
       "1788   2015-10-21 00:44:00\n",
       "1694   2015-10-21 00:44:00\n",
       "1695   2015-10-21 00:44:00\n",
       "1797   2015-10-21 00:44:00\n",
       "1796   2015-10-21 00:44:00\n",
       "               ...        \n",
       "1534   2015-10-21 00:25:00\n",
       "1535   2015-10-21 00:25:00\n",
       "1537   2015-10-21 00:25:00\n",
       "1538   2015-10-21 00:25:00\n",
       "1621   2015-10-21 00:17:00\n",
       "1880   2015-10-21 00:05:00\n",
       "1620   2015-10-21 00:04:00\n",
       "1539   2015-10-21 00:02:00\n",
       "1520   2015-10-21 00:02:00\n",
       "1540   2015-10-21 00:02:00\n",
       "1536   2015-10-21 00:02:00\n",
       "1598   2015-10-21 00:00:00\n",
       "1594   2015-10-20 23:47:00\n",
       "1593   2015-10-20 23:47:00\n",
       "1596   2015-10-20 23:47:00\n",
       "1622   2015-10-20 13:25:00\n",
       "1663   2015-10-20 13:25:00\n",
       "1523   2015-10-20 06:30:00\n",
       "1517   2015-10-20 06:30:00\n",
       "1521   2015-10-20 06:30:00\n",
       "1522   2015-10-20 06:30:00\n",
       "1525   2015-10-20 06:30:00\n",
       "1524   2015-10-20 06:30:00\n",
       "1528   2015-10-20 06:30:00\n",
       "1527   2015-10-20 06:30:00\n",
       "1526   2015-10-20 06:29:00\n",
       "1542   2015-10-20 06:29:00\n",
       "1519   2015-10-20 06:29:00\n",
       "1529   2015-10-20 06:29:00\n",
       "1530   2015-10-20 06:29:00\n",
       "Name: dt_mod, dtype: datetime64[ns]"
      ]
     },
     "execution_count": 96,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# validate last seen for hdfs user\n",
    "df[(df.user=='hdfs')]['dt_mod'].sort_values(ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    1708.000000\n",
       "mean        2.074941\n",
       "std         0.770885\n",
       "min         2.000000\n",
       "25%         2.000000\n",
       "50%         2.000000\n",
       "75%         2.000000\n",
       "max        10.000000\n",
       "Name: num_replicas, dtype: float64"
      ]
     },
     "execution_count": 97,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# average replicas\n",
    "df['num_replicas'].describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "user       group \n",
       "ambari-qa  hadoop    2.000000\n",
       "           hdfs      5.047619\n",
       "anonymous  hdfs           NaN\n",
       "hbase      hdfs      2.000000\n",
       "hcat       hdfs           NaN\n",
       "hdfs       hadoop    2.000000\n",
       "           hdfs      2.000000\n",
       "hive       hdfs           NaN\n",
       "mapred     hadoop         NaN\n",
       "           hdfs           NaN\n",
       "oozie      hadoop         NaN\n",
       "           hdfs      2.000000\n",
       "spark      hdfs           NaN\n",
       "yarn       hadoop         NaN\n",
       "Name: num_replicas, dtype: float64"
      ]
     },
     "execution_count": 98,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# average replica user:group\n",
    "df.groupby(['user','group'])['num_replicas'].mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "is_dir  permissions\n",
       "False   -r--r--r--       10\n",
       "        -rw-r-----       49\n",
       "        -rw-r--r--     1462\n",
       "        -rwx--x---        4\n",
       "        -rwxr-xr-x      143\n",
       "        -rwxrwx---       40\n",
       "True    dr-xr-xr-x        8\n",
       "        drwx------       22\n",
       "        drwx--x--x        2\n",
       "        drwx-wx-wx        1\n",
       "        drwxr-xr-x      941\n",
       "        drwxrwx---       47\n",
       "        drwxrwxr-x        2\n",
       "        drwxrwxrwt        1\n",
       "        drwxrwxrwx        5\n",
       "Name: fname, dtype: int64"
      ]
     },
     "execution_count": 99,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# permission count by object type\n",
    "df.groupby(['is_dir','permissions'])['fname'].count()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x10a35d198>"
      ]
     },
     "execution_count": 100,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAf4AAAEsCAYAAAAimdEBAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzs3XlclOX+//HXsNzSKKCASpzAfcnUSgyzQk0o7bSaKZmW\nljtKfXPBNjPNEpOyQkWPUErm0jmpLZ5zLDl4DDMNyU5mHVNUMBcWZRURmPv3hz/mODHgIDAzl36e\nj4cPupe5531f0Fxz39dyG3Rd1xFCCCHENcHF0QGEEEIIYT9S8QshhBDXEKn4hRBCiGuIVPxCCCHE\nNUQqfiGEEOIaIhW/EEIIcQ2Ril+IOsrNzeXZZ5+lbdu2NGnSBB8fH/r27UtaWlqtr8vOziYoKIhx\n48bZKakAaNu2Lffcc0+19a+99houLi6cOHHCAamEcBw3RwcQQiW5ubncdtttnD59mieffJKePXtS\nXFzMnj17uP7662t9bWlpKbm5uVLR2JnBYMBgMDg6hhBOQyp+Iepg8eLFHDt2jJSUFPr371+n17Zp\n04bff/8dT0/PRkonrJE5yoSwJBW/EHWQnp5Oq1at6lzpV2nRokUDJxJCiLqRNn4h6sDf35+8vDwy\nMjJs2v/8+fO4uLhY/Bs/fnyN+48ZM6ba/pf+s9ZMUFZWxmuvvUb79u3x8PDgxhtvZOnSpVd8jpfa\nvn07Li4ufPzxxyQmJtKjRw+MRiM9evTgyy+/rPF1//znP7nzzjtp2rQpLVu2ZOzYseTk5NS4f9u2\nbXn66ac5ffo0w4YNw9vbG09PTwYMGNAg51FXmZmZPPvss3Tq1Amj0Yivry933HEH//73v63uX9ff\nQVW5/vvf/+af//wnt9xyCx4eHrRu3Zq4uLjGOi0hALniF6JOnn76aZKSkrjnnntISkrizjvvrHV/\nTdNYuXIlBoMBXdcZP358re3NY8eOpV+/fhbr8vLyeOGFF3jiiScICAiw2KbrOkOGDCElJYWnn36a\nG2+8kdTUVKKiosjMzGThwoVXfrKXeOuttygoKGDChAm4urqydOlSHn30UVJTUwkJCbHYd+3atTz5\n5JP079+fefPmcfbsWZYuXUpqairff/89Xl5e1Y5vMBgoLS1l4MCBtGzZkvnz5zusWeTEiRP06tUL\nd3d3nnzySdq1a0dxcTH79u3Dx8en2v71+R0kJyfzzjvvMGbMGMaNG8euXbu46667GvP0hABdCFEn\nK1as0D08PHSDwaDff//9elpams2vNRgM+vjx4+v0fiNGjND9/f31M2fOVNv24Ycf6q6urvqOHTss\n1s+ZM0d3cXHRf/nllzq91x+lpKToBoNBv+mmm/SzZ8+a1x85ckRv0qSJ/sgjj1jsX1BQoHt5eekT\nJkywWH/gwAHdaDTqL774otX3adOmje7u7q5PmjSpXnlrOvY999xTbf2cOXN0g8Gg//777xbr33nn\nHd1gMOi//vqrTce/kt9BVbk2a9ZM3759ex3ORoj6k1v9QtTRhAkTOHjwIJMnT2bbtm2EhITw0ksv\nNcp7bdq0ifXr17Ns2TKr/QNWr15Nly5d6Ny5M6dOnTL/e+yxx9B1nc8//7xBcsyaNYvmzZubl9u2\nbcuQIUP44osvKCsrs8hbVFTEsGHDLPK0aNGC22+/nc2bN9f4Hl27dmXJkiUNkrc+rrvuOuDi8Etb\n1Od38Morr1xxfxEhrpTc6hfiCgQGBrJ06VKmTZvGM888Q0xMDP7+/jz77LMN9h55eXlMnjyZiIgI\nhgwZYnWfH3/8kfz8fKtDCQ0GA5mZmQ2SxdXVtdq63r17s2HDBo4fP06HDh3MeQDuvfdeq8cxGo1W\n1xsMBnr37m31feztySefZPny5QwcOJCwsDAeeeQRHnrooWrNLFXq8zvo27dvg+UWwlZS8QtRDx06\ndODrr7+me/fuzJ07l6ioqAYbMz516lSAWq+Ci4qK6N+/Py+++KLV7YGBgQ2SxRo/Pz8AcnJyzBV/\nUVERAJ9//jmaplV7jYuL899kbNq0Kbt372blypV88sknTJ06lSlTpvDwww+zYsUKWrZsabG/I38H\nQlwJqfiFqCdN03jggQd49913OXPmDL6+vvU+5saNG9mwYQOffPJJrcfz8fHh/PnzNV5hN6aqW+He\n3t7mdVXNEZ06daJLly52z2SNpmkUFxdXW3/hwgXz9j9q0qQJU6dOZerUqZw6dYr4+HgWLFjA0KFD\n2bFjh8W+jvwdCHElnP/rtxAKyM/PR9M0i0rwSuXm5jJ58mSGDh3KY489Vuu+ISEh7N27l0OHDtX7\nfWuTn59fbd23336L0WikU6dO5nV9+vQBYN26dY2apy5at27NkSNHqq0/duwYmqZddm4Ff39/5s6d\ny5gxY0hNTa1WFvb6HQjRUKTiF6IOysvLq607cuQIn376KYMGDcLNrf430aZOnYqu6yxbtuyy+0ZG\nRlJRUcGUKVOsZmsob775JsePHzcv//LLL3z55Zc88sgjFuf84IMPcsMNN/DOO++wf//+RstTF6Gh\noZw+fZpPP/3UvC47O5t//OMfBAcH29SvoLKykkOHDnHdddfRtGlTi232+h0I0VDkVr8QdfDee+/x\n4Ycfmjt7ZWRk8MEHH9CsWTPefffdeh9/48aNfPLJJzzyyCN89tln1baPGDHCouK57777+L//+z/e\nffddbr31VkaNGoXRaCQzM5Pt27eTkpLSIGPh27Zty2233UZkZCSurq7ExcVhNBqZO3euxX6aprFu\n3Truu+8++vbty7hx4+jUqRN5eXns2bOHIUOG8Mwzz1Q7vt6I0+pOnTqV5cuXM2rUKP7+97/j7e3N\np59+SkFBATNmzKi2/969e5k8eTKDBg3C39+fs2fPsnHjRvbt28err76Ku7u7xf72+h0I0VDsXvGb\nTCY+/vhjvvvuOyoqKujTp495tjKAQ4cOkZCQQE5ODt7e3owZM4aePXuaX19WVkZiYiJ79+7Fzc2N\n0NBQRo4cKQ/hEHYREhLCN998w5o1a8jOzqZ169Y89dRTvPrqq9U6fV2JpUuXYjAY2Lx5c7WhbwaD\ngXvvvbfaFec777xDSEgIS5cu5Y033kDXdf70pz8xcOBAKioq6p0JLl7VlpWV8dZbb3H06FFuueUW\n3n33XXOnvkvdeeedpKWl8frrr7N+/Xpzv4dbb72V9u3bWz1+Y/7/GxAQwI4dO3jhhRf429/+xoUL\nF+jWrRuLFi2yOlqiadOm3HDDDaxatYqcnBxatGhBjx49+Oyzz3jwwQetvseV/A7kM0s4ikFvzK/a\nVnzxxRfs27ePmTNn4u7uTnx8PNdffz1Dhw6lvLycyMhIIiIiCA8P58iRI8TExLBo0SLzbF+rVq0i\nKyuLWbNmARAbG0uvXr0YPHiwPU9DiGvC9u3bGThwIGvWrOGJJ55wdBwhRAOwexv/L7/8Qt++ffHw\n8MDV1ZUxY8awdetWTCYTu3fvRtM0wsPDAWjXrh0hISGkpKQAF9tXt2/fztChQ9E0DU3TiIiIYOvW\nrfY+DSGEEEJJdq/4O3TowM6dOykuLsZkMnHixAnKyso4c+YMWVlZBAUFkZaWxujRo0lPTycwMJCs\nrCzgYm/n0tJS/P39iYqKIi4ujsDAQE6ePInJZLL3qQghhBDKsXsb/5AhQ6ioqGD27NmYTCZuvPFG\n4OIkGEVFRRiNRvMDTQwGA5qmUVhYaN4HLs7+pes6uq6jaRq6rlNUVNQgQ6mEEJakLVqIq4vdK34X\nFxciIiKIiIgALk6ikZKSQsuWLfH09CQ/P5/g4GCSkpIA2LJli7lHbNXPyspK82xmJSUlFtuEEA1n\nwIABVFZWOjqGEKIBOXw433fffUfbtm1p1qwZQUFBfPvttxbbs7KyaNOmDXBxilAPDw+OHTtGt27d\nzNsDAgJqnQo0OTm58U5ACCGEcEJhYWFW19u94j9//jwmkwmj0ciJEyfYsGED48ePBy7O+pWUlERy\ncjJhYWFkZGSQnp5u7k3s7u7OwIED2bhxo3kYUdXEKZfTq1evxjspIYQQwomkp6fXuM3unfsyMzOZ\nMWMGkyZNYtGiRYwYMYJbbrkFADc3N6Kjo9m2bRtjx45lyZIlREZGmofywcUJTHx8fIiMjCQqKoqg\noCCbKv7GkJqa6pD3vVKq5QXJbA+q5QXJbA+q5QXJbCu7X/F37ty51qlIO3TowIIFC2rcrmkakZGR\njRFNiKvGmQvnOHuh1KZ9je0DOFycZ/OxW2jX4aNZf7yuEML52X0CH0dITk6WW/3imnK4OI+VR/c0\nyrHHtw2hQ7P6P4FQCNF40tPTa2zjl4f0CCGEENcQqfjrQbX2JNXygmQW1qlYxqplVi0vSGZbScUv\nhBBCXEOkjV+Iq5C08QtxbZM2fiGEEEIAUvHXi2rtSarlBcksrFOxjFXLrFpekMy2kopfCCGEuIZI\nG78QVyFp4xfi2lZbG79DHtKzefNmvv76a0wmE127dmXs2LE0a9YMgEOHDpGQkEBOTg7e3t6MGTOG\nnj17ml9bVlZGYmIie/fuxc3NjdDQUEaOHCmPDhVCCCFsYPdb/WlpaSQnJ/Pmm2+ybNkyPD09SUhI\nAKC8vJyFCxcSHh5OYmIiUVFRLF26lMLCQvPr161bR15eHvHx8cTFxZGZmcnWrVvtfRqAeu1JquUF\nySysU7GMVcusWl6QzLaye8Wfm5tLYGAg3t7eGAwGunfvTnZ2NgC7d+9G0zTCw8MBaNeuHSEhIaSk\npAAXvxhs376doUOHomkamqYRERHhsIpfCCGEUI3dK/4BAwZQUVHBqlWr2LJlC5s2beKpp54CICsr\ni6CgINLS0hg9ejTp6ekEBgaSlZUFXPzSUFpair+/P1FRUcTFxREYGMjJkycxmUz2PhXuuusuu79n\nfaiWFySzsE7FMlYts2p5QTLbyiFt/H5+fvz++++cO3eOJk2a4OJy8ftHUVERRqMRg8GArusYDAY0\nTTPf6i8qKgLAaDSi6zq6rqNpGrquU1RUhLe3tyNORwghhFCG3Sv+Dz74AE9PT15++WUAMjMzmTdv\nHnFxcXh6epKfn09wcDBJSUkAbNmyBU9PTwDzz8rKSpYsWQJASUmJxbaapKammr9ZVbWp1He5al1D\nHa+xl1XLe2lWZ8ljy3J8fDw9evRwaB5j+wAaS0FhAan7fnHo+f30009MnjzZYe9/JctV65wlz9WW\nVz4vLJeNxpofnW334XwzZsxg5MiR3HrrreZ1Y8eOZc6cOWRlZbF+/Xri4uLM25YvX05AQAAPPfQQ\n5eXljBs3jlmzZtGtWzcAfv31V1asWMHixYtrfM/GGs536ZcJFaiWFyTzlbrah/M5QxnXlWqZVcsL\nkvlSTjVlb7du3fjyyy8pLCzEZDLxj3/8Azc3NwICAggJCeHChQskJycDkJGRQXp6OgMGDADA3d2d\ngQMHsnHjRsrKyigrK+PTTz9l0KBB9j4NQL32JNXygmQW1qlYxqplVi0vSGZb2f1W/6hRo/j444+J\njo7mwoULtGvXjpdffhk3t4tRoqOjSUhIYO3atXh7exMZGYmXl5f59SNGjCAhIYHIyEhcXV0JDQ11\nWMUvhBBCqEZm7qsH1W4rqZYXJPOVklv9zke1zKrlBcl8Kae61S+EEEIIx5ErfiGuQlf7Fb8QonZy\nxS+EEEIIQCr+elFtXmjV8oJkFtapWMaqZVYtL0hmW0nFL4QQQlxDpI1fiKuQtPELcW2TNn4hhBBC\nAFLx14tq7Umq5QXJLKxTsYxVy6xaXpDMtrL7zH1r167lm2++MS+bTCby8/NZvnw5LVq04NChQyQk\nJJCTk4O3tzdjxoyhZ8+e5v3LyspITExk7969uLm5ERoaysiRIzEYDPY+FSGEEEI5Dm/j37FjB9u2\nbWPevHmUl5cTGRlJREQE4eHhHDlyhJiYGBYtWmSetnfVqlVkZWUxa9YsAGJjY+nVqxeDBw+u8T2k\njV9ca6SNX4hrm1O38X/++ec89NBDAOzevRtN0wgPDwegXbt2hISEkJKSAkB5eTnbt29n6NChaJqG\npmlERESwdetWh+UXQgghVOLQij89PZ3Kykp69+4NQFZWFkFBQaSlpTF69GjS09MJDAwkKysLgNzc\nXEpLS/H39ycqKoq4uDgCAwM5efIkJpPJ7vlVa09SLS9IZmGdimWsWmbV8oJktpVDK/7PPvvMfLUP\nUFRUhNFoxGAwoOs6BoMBTdMoLCw0bwcwGo3ouo6u62iahq7r5m1CCCGEqJndO/dVOXjwIKdOnaJf\nv37mdV5eXmRmZhIcHExSUhIAW7ZswdPTE8D8s7KykiVLlgBQUlJisc2eVHsKlGp5QTIL61QsY9Uy\nq5YXJLOtHHbF/9lnn3H//ffj6upqXnfpbf0qWVlZtGnTBgA/Pz88PDw4duyYxfaAgABcXGo/lUtv\np6SmpsqyLF/VywWFBTSWgsICh5+fLMuyLNu2bI1DevX//vvvvPLKKyxbtozrrrvOvL6iooIpU6Yw\nfPhwwsLCyMjIICYmhtjYWHOv/tWrV5OVlcXMmTOBi736g4ODHdKrPzVVrWc/q5YXJPOVutp79TtD\nGdeVaplVywuS+VK19ep3yK3+L774grCwMItKH8DNzY3o6GgSEhJYu3Yt3t7eREZGmit9gBEjRpCQ\nkEBkZCSurq6EhoYyaNAge5+CEEIIoSSHj+O3BxnHL641V/sVvxCidg12xa/rOllZWeTl5VFWVoav\nry833HBDtSt3IYQQQjgnmyr+4uJiNm/ezDfffEN+fr7lAdzcCA4O5sEHH6RTp06NEtJZqdaepFpe\nkMzCOhXLWLXMquUFyWyry1b8aWlp/OUvf6F9+/aMHDmSzp074+Xlhbu7O4WFhfz+++/88MMPLFiw\ngH79+vHkk09a9NQXQgghhPOotY1/165drFmzhsmTJ9O9e/daD1RQUEBCQgIAzz///GWH19mTtPGL\na4208Qtxbbuiufp1XWf//v3MnTv3spU+gLe3N9OmTeP666+vNhZfCCGEEM6hxorfYDAwfvx4/Pz8\nbD6YwWDgiSeeME+4c7W73CQJzka1vCCZhXUqlrFqmVXLC5LZVs5zP14IIYQQjU7G8QtxFZI2fiGu\nbfUax5+fn8/+/fsxmUyEhITg4eFBbm4ua9eu5ciRI7Rs2ZIhQ4Zw4403NnhwIYQQQjSsWm/1Hz16\nlOnTp7NixQoSExOZOXMmubm5zJs3j4KCAvr27Yuu68yfP58jR47Y/KbZ2dm8+eabjB8/nkmTJrF2\n7VrztkOHDvHCCy8wduxYpk2bxn/+8x+L15aVlbFs2TLGjh3LxIkTWbNmDY66aaFae5JqeUEyC+tU\nLGPVMquWFySzrWq94l+3bh1du3Y1D89LSkpi/vz5tG3blmnTppn3e//999mwYQMvvPDCZd+wsrKS\nt956i8cff5zevXsDcP78eQDKy8tZuHAhERERhIeHc+TIEWJiYli0aJF5vv5169aRl5dHfHw8cPEh\nPVu3bq31IT1CCCGEuKjWK/5Dhw7x4IMP4ubmhouLC48++ignT55k4MCBFvsNHDiQ3377zaY33Llz\nJ507dzZX+gAeHh4A7N69G03TCA8PB6Bdu3aEhISQkpICXPxisH37doYOHYqmaWiaRkREBFu3brX9\njBuQajNEqZYXJLOwTsUyVi2zanlBMtuq1oq/uLiY5s2bm5e9vLxwc3OrNsTPx8eH4uJim97wwIED\nFBYW8vLLLxMVFcXbb79Nbm4uAFlZWQQFBZGWlsbo0aNJT08nMDDQPC9Abm4upaWl+Pv7ExUVRVxc\nHIGBgZw8eRKTyVSnExdCCCGuRZcdznfu3DmKi4spLi6mqKgIuHhrvmpdcXExpaWlNr/hmTNnqKio\nYNq0abz//vt069aNBQsWYDKZKCoqwmg0YjAY0HUdg8GApmkUFhYCmN/faDSi6zq6rqNpGrqum7fZ\nk2rtSarlBcksrFOxjFXLrFpekMy2umyv/hdffLHaupdffvmK39DLy4vWrVvj63txONB9993HJ598\nwsmTJ/Hy8iIzM5Pg4GCSkpIA2LJlC56engDmn5WVlSxZsgSAkpISi21CCCGsO3PhHGcv2HahZmwf\nwOHiPJv2baFdh49mrE80YUe1VvyvvvqqzQcyGAw27dexY0e++eYbhg0bZhnEzY3AwEB27txpsT4r\nK8s8E6Cfnx8eHh4cO3aMbt26mbcHBARc9tkAlz4Bqeobliw7//Jdd93lVHlsWa5a58g8xvYBNJaC\nwgJS9/3i8PKu4ujft0rLZy+U1m1+hzO27Tbcpws+Ae0dfn7yefG/ZaOx5i9idp/Ap6ysjKlTpzJq\n1Cj69+9PamoqmzdvJjY2lvLycqZOncrw4cMJCwsjIyODmJgYYmNjzb36V69eTVZWFjNnzgQu9uoP\nDg6utVe/TOAjrjUygY+wprH+LuRvwvlc0UN6GkuTJk14+eWX+eqrr5g4cSL//Oc/ef755wFwd3cn\nOjqabdu2MXbsWJYsWUJkZKS50gcYMWIEPj4+REZGEhUVRVBQEIMGDbL3aQDqtSeplhcks7BOxTJW\nMbNqVCxjp2vjnzt3bp0ONmfOHJv2a9u2LW+88YbVbR06dGDBggU1vlbTNCIjI+uUSwghhBAX1Vrx\nHzhwABcXF3r06IG/vz9AjbPk2drGfzVRbcyoanlBMgvrVCxjFTOrRsUydkTmWiv+N954g507d7Jr\n1y5KSkq48847ueOOOyzG9gshhBBCHbW28Xfs2JHRo0ezbNkynnjiCY4fP8706dP529/+Zq98Tk21\n9iTV8oJkFtapWMYqZlaNimXsiMw2de67cOECubm55OTkADJmXgghhFBVrbf6v/vuO3bu3Mn+/fu5\n+eabuffee7n11ltxc7vsvD/XBNXak1TLC5JZWKdiGauYWTUqlrHTtfEvXrwYFxcXunTpwvnz50lO\nTuZf//oX8L9OfpdOrztr1qzGTyyEEEKIK1brrf5+/fpx11130apVK7y8vPD09KRZs2Y0a9YMT09P\n83LVf19rVGtPUi0vSGZhnYplrGJm1ahYxk43jn/KlCn2yiGEEEIIO7D7zH1XE9Xak1TLC5JZWKdi\nGauYWTUqlrHTtfE3loiICHx8fMzLQ4YM4d577wXg0KFDJCQkkJOTg7e3N2PGjKFnz57mfcvKykhM\nTGTv3r24ubkRGhrKyJEjr8kJhIQQQoi6qvGKX9d1Vq5cyZkzNj6e6f9bt24dx44du+x+8fHx5n9V\nlX55eTkLFy4kPDycxMREoqKiWLp0KYWFhRbHz8vLIz4+nri4ODIzM9m6dWudMjYU1dqTVMsLkllY\np2IZq5hZNSqWsVON4zcYDNx0003Mnj2bAwcOXPZAhYWFvP322/z+++8EBgZeUZjdu3ejaRrh4eEA\ntGvXjpCQEFJSUoCLXwy2b9/O0KFD0TQNTdOIiIhwWMUvhBBCqKbWW/133HEHmqaxePFiOnXqRN++\nfencuTPe3t64ublRVFTEyZMn2bt3L9u3byc0NJRRo0bh4lJ71wFXV1eioqLQNI0777yTRx55BBcX\nF7KysggKCiItLY24uDiee+45AgMDOXjwIAC5ubmUlpbi7+9PVFQUnTt3ZuLEiZw8eRKTyXTZ921o\nqrUnqZYXJLOwTsUyVjGzalQsY6ds4+/duzddunRh8+bNrFmzhvz8fMsDuLlx6623MmvWLDp37mzT\nmyYkJNCkSRNKSkqIjY3F1dWVhx9+mKKiIoxGo8XcAJqmmW/1FxUVAWA0GtF1HV3X0TQNXdcpKirC\n29u7rucvhBBCXFNs6tzn6enJk08+yciRI8nKyiI3N5eysjJ8fX254YYbaNq0aZ3e1Gg0AuDl5cVD\nDz3Ep59+ysMPP4ynpyf5+fkEBweTlJQEwJYtW8xzBFT9rKysZMmSJQCUlJRYbKtJamqq+ZtVVZtK\nfZer1jXU8Rp7WbW8l2Z1ljy2LMfHx9OjRw+H5jG2D6CxFBQWkLrvF4ee308//cTkyZMd9v5Xsly1\nztF5GkNBYQE087VLfvm8sG25qp61xqDX9JxdO/nuu+/4xz/+wdy5c9m5cyfr168nLi7OvH358uUE\nBATw0EMPUV5ezrhx45g1axbdunUD4Ndff2XFihUsXry4xvdITk6mV69eDZ790i8TKlAtL0jmK3W4\nOI+VR/c0yrHHtw2hw///kHcUZyjjunKGzI31d+EMfxPgHGVcV42VOT09nbCwMKvb7D6O/9y5c+Zb\n9oWFhXzxxRfmznwhISFcuHCB5ORkADIyMkhPT2fAgAEAuLu7M3DgQDZu3EhZWRllZWV8+umnDBo0\nyN6nAajXnqRaXpDMwjoVy1jFzKpRsYydso2/oWVkZLBixQrKy8sxGo088MADhIaGAhcr9ujoaBIS\nEli7di3e3t5ERkbi5eVlfv2IESNISEggMjISV1dXQkNDHVbxCyGEEKqxe8XfvXt3i1v5f9ShQwcW\nLFhQ43ZN04iMjGyMaHWm2m0l1fKCZBbWqVjGKmZWjYpl7IjMMmWvEEIIcQ2Rir8eVPtmqVpekMzC\nOhXLWMXMqlGxjB2RWSp+IYQQ4hoiFX89qDYvtGp5QTIL61QsYxUzq0bFMnaqufqFEEIIcfWRir8e\nVGtPUi0vSGZhnYplrGJm1ahYxtLGL4QQQohGJRV/PajWnqRaXpDMwjoVy1jFzKpRsYwdkdnuE/hc\nauHChaSnpxMfH4+Pjw8Ahw4dIiEhgZycHLy9vRkzZgw9e/Y0v6asrIzExET27t2Lm5sboaGhjBw5\nEoPB4KjTEEIIIZThsCv+PXv2YDKZLNaVl5ezcOFCwsPDSUxMJCoqiqVLl5ofywuwbt068vLyiI+P\nJy4ujszMTLZu3Wrv+IB67Umq5QXJLKxTsYxVzKwaFctYiTb+X3/9lW3btlFcXGyx/vz58zYf4/z5\n83z88ceMHTvWYv3u3bvRNM380J527doREhJCSkoKcPGLwfbt2xk6dCiapqFpGhEREQ6r+IUQQgjV\n2FzxFxYWMmfOHObPn8/KlSstrsI/++yzWuff/6NPPvmEfv360apVK4v1WVlZBAUFkZaWxujRo0lP\nTycwMJCsTXFEAAAgAElEQVSsrCwAcnNzKS0txd/fn6ioKOLi4ggMDOTkyZPV7h7Yg2rtSarlBcks\nrFOxjFXMrBoVy9ipx/F/8MEHaJpGQkICbm6WXQM6d+7MwYMHbTrOsWPH2LdvHw8//HC1bUVFRRiN\nRgwGA7quYzAY0DTN/CWj6nG+RqMRXdfRdR1N09B13bxNCCGEEDWzuXPfDz/8wOzZs/Hw8Ki2zdPT\nk5KSksseQ9d1Vq5cyZgxY3Bzc0PXdYvtXl5eZGZmEhwcTFJSEgBbtmzB09PT/D4AlZWVLFmyBMD8\nvlXb7Em19iTV8oJkFtapWMYqZlaNimXs1G38Li4uVFRUWN124MABfH19L3uMrKwsjh49yuLFi3n6\n6ad55plnAJg2bVq12/qXvqZNmzYA+Pn54eHhwbFjxyy2BwQE4OJS+6lcejslNTVVlmX5ql4uKCyg\nsRQUFjj8/GT5ypcbw6V/b44+P1m+/O/boP/xsrsGy5Yto6ioiOnTpzN69GgWLVrE9ddfz44dO0hM\nTOTBBx9k2LBhthzKQkREhHk4X0VFBVOmTGH48OGEhYWRkZFBTEwMsbGxeHl5AbB69WqysrKYOXMm\nALGxsQQHBzN48OAa3yM5OZlevXrVOdvlpKaq9exn1fKCZL5Sh4vzWHl0T6Mce3zbEDo0u/wX/cbk\nDGVcV86QubH+LpzhbwKco4zrqrEyp6enExYWZnWbzbf6R48eTUxMDOPHj6eiooK4uDjy8vIoKCjg\ntttuY8iQIfUO6ubmRnR0NAkJCaxduxZvb28iIyPNlT7AiBEjSEhIIDIyEldXV0JDQxk0aFC931sI\nIYS4Fth8xQ9gMpn44YcfOHDgACUlJXh5eXHzzTdz0003NWbGemusK34hnNXVfsUvrszVfsUv/qdB\nrvjhYjt/cHAwwcHBDRJMCCGEEPYlc/XXQ2N3mGloquUFySysU7GMVcysGhXL2BGZ6zSBT01MJhO5\nubkNEkgIIYQQjcfmin/ChAl89NFHlJeXV9uWnZ3NlClTGjSYClTrPapaXpDMwjoVy1jFzKpRsYyd\nehy/ruvs3buX6Ohom2fpE0IIIYRzqVMbf3R0NP379+f111+v8er/WqJae5JqeUEyC+tULGMVM6tG\nxTJ26jZ+AFdXVx555BHefPNNDhw4IFf/QgghhGLqNJyvSmBgIPPnz2fTpk3MmzeP2267raFzKUG1\n9iTV8oJkFtapWMYqZlaNimXs1G38f+Tq6spjjz3G/PnzOX78eENmEkIIIUQjsbnif/XVV2nRokW1\n9W3btmXBggU8++yzNh3nxIkTzJgxg8jISKKioli7di0mk8m8/dChQ7zwwguMHTuWadOm8Z///Mfi\n9WVlZSxbtoyxY8cyceJE1qxZU+0pf/aiWnuSanlBMgvrVCxjFTOrRsUydkRmm2/11zYtr5ubG3fe\neadNx2ndujUxMTG4ublx/vx55s6di7+/PwMHDqS8vJyFCxcSERFBeHg4R44cISYmhkWLFpnn61+3\nbh15eXnEx8cDFx/Ss3Xr1lof0iOEEEKIi+w+c5+rqytubm7ouk5mZibFxcXmx+7u3r0bTdMIDw8H\noF27doSEhJCSkgJAeXk527dvZ+jQoWiahqZpREREsHXrVnufBqBee5JqeUEyC+tULGMVM6tGxTJW\nqo2/PvLy8njqqad47bXXGDZsGB06dAAgKyuLoKAg0tLSGD16NOnp6QQGBpKVlQVAbm4upaWl+Pv7\nExUVRVxcHIGBgZw8edKiuUAIIYQQ1jmk4vf19eWjjz7ihRde4KOPPmL//v0AFBUVYTQaMRgM6LqO\nwWBA0zTzdMFFRUUAGI1GdF1H13U0TUPXdfM2e1KtPUm1vCCZhXUqlrGKmVWjYhk7XRv/nj178PT0\n5MYbb+TAgQOXPVi3bt3q9OY9e/bkvvvu46uvvqJ79+54enqSn59PcHAwSUlJAGzZsgVPT08A88/K\nykqWLFkCQElJicW2mqSmpppvqVQVdH2XLz12QxyvsZdVy6vq8k8//eTwPMb2ATSWgsICUvf94tDz\n++mnn5zm963a/3+NoaCwAP7/Y3kdfX6qLTfW54XRaKQmBr2WLvFjx47lhhtuYO7cuURERNR4kCob\nNmy47D5/tGnTJnJycpgwYQI7d+5k/fr1xMXFmbcvX76cgIAAHnroIcrLyxk3bhyzZs0yf8n49ddf\nWbFiBYsXL67xPZKTk+nVq1edswmhqsZ67jrIs9dV1lh/F/I34XzS09MJCwuzuq3WK/6FCxfSpEkT\n8/Kbb75Jy5Yt6xXm9OnTNGvWjKZNm3L06FG++uorZsyYAUCfPn1ISkoiOTmZsLAwMjIySE9P54kn\nngDA3d2dgQMHsnHjRnO/gE8//ZRBgwbVK5MQQghxrai1jd/Pz8/iFnqzZs3w8vKq8Z8tfvnlF6Kj\no5k8eTIJCQlMmTLFXIm7ubkRHR3Ntm3bGDt2LEuWLCEyMtLi2CNGjMDHx8c8D0BQUJDDKn7V2pNU\nywuSWVinYhmrmFk1Kpax07XxX+q+++6rtc3AVgMGDGDAgAE1bu/QoQMLFiyocbumaURGRtY7hxBC\nCHEtqrWN/2ohbfziWiNt/MIaaeO/dtTWxu+Q4XxCCCGEcAyp+OtBtfYk1fKCZBbWqVjGKmZWjYpl\n7NRt/H909OhRTpw4QatWrejYsWNDZhJCCCFEI6m14l+7di29evWia9euFutXrVrFP/7xD1xcXDCZ\nTPTs2ZPo6Gjc3d0bNayzUW1eaNXygmQW1qlYxipmVo2KZex0c/WnpKRUmwP/3//+N1u3buW5557j\n448/JjY2ltzcXD7//PNGDSqEEEKI+qu14j937pzFGHqTycQnn3zCAw88wB133IGLiwuBgYGMGDGC\nHTt2NHpYZ6Nae5JqeUEyC+tULGMVM6tGxTJ2RObLTuBT9WQ8uDgsrri4mIcffthiP39/f3Jzcxsn\noRBCCCEaTK1t/HfddRdr1qzBzc2N0tJSPvroI/785z/TrFkzi/2ys7Np3ry5TW+Ynp7O6tWrKSsr\nw9XVlbCwMB599FHz9kOHDpGQkEBOTg7e3t6MGTOGnj17mreXlZWRmJjI3r17cXNzIzQ0lJEjR2Iw\nGOpy3g1CtfYk1fKCZBbWqVjGKmZWjYpl7IjMtVb8Dz/8MMeOHSM2NhaA/v37M2zYsGr7HTp0iFtv\nvdWmN+zUqRMLFy7Ew8ODU6dOMXv2bHx9fenfvz/l5eUsXLiQiIgIwsPDOXLkCDExMSxatMjc5LBu\n3Try8vKIj48HIDY2lq1btzJ48OA6nbgQQghxLar1Vr+macyYMYMPPviA1atXExkZiaura7X9Hn/8\nccaNG2fTG3p6euLh4QFcbCLo1q0bhw8fBmD37t1omkZ4eDgA7dq1IyQkhJSUFADKy8vZvn07Q4cO\nRdM0NE0jIiKCrVu32n7GDUi19iTV8oJkFtapWMYqZlaNimXsdG38VZo2bWqurBvaiRMnCAoKAiAr\nK4ugoCDS0tIYPXo06enpBAYGmvsZ5ObmUlpair+/P1FRUcTFxREYGMjJkyerjT4QQgghRHUOnblv\nz549nD9/nn79+gFQVFSE0WjEYDCg6zoGgwFN0ygsLDRvBzAajei6jq7raJqGruvmbfakWnuSanlB\nMgvrVCxjFTOrRsUydro2/sZ05swZPvroI6ZPn46maQB4eXmRmZlJcHAwSUlJAGzZssX8aOCqn5WV\nlSxZsgSAkpISi21CCCGEqJlDKv7CwkIWLVrEpEmTaNu2rXl9YGAgO3futNg3KyuLNm3aABeHF3p4\neHDs2DG6detm3h4QEICLS+03L1JTU83frKraVOq7XLWuoY7X2Muq5b00q7PksWU5Pj6eHj16ODSP\nsX0AjaWgsIDUfb849Px++uknJk+e7LD3v5LlqnWOztMYCgoL4P8/nU8+L5zj88JoNFITuz+Wt7i4\nmIULFzJs2DCLYXoAFRUVTJkyheHDhxMWFkZGRgYxMTHExsaae/WvXr2arKwsZs6cCVzs1R8cHFxr\nr/7GeizvpV8mVKBaXpDMV+pqfyyvM5RxXTlD5qv9sbzOUMZ11ViZa3ssr90r/pUrV/Lvf//b4tZ8\nx44dmT59OgCHDx8mISGB7OxsvL29eeqpp7jlllvM+164cIGEhAT27t2Lq6sroaGhjBo1qtZx/I1V\n8QvhrK72il9cmau94hf/U1vFb/db/ePHj2f8+PE1bu/QoQMLFiyocbumaURGRjZGNCGEEOKq59Be\n/apTbcyoanlBMgvrVCxjFTOrRsUydtpx/EIIIYS4OkjFXw+qdSJRLS9IZmGdimWsYmbVqFjGjsgs\nFb8QQghxDZGKvx5Ua09SLS9IZmGdimWsYmbVqFjG0sYvhBBCiEYlFX89qNaepFpekMzCOhXLWMXM\nqlGxjKWNXwghhBCNSir+elCtPUm1vCCZhXUqlrGKmVWjYhk7IrNDHtJz8OBBNm/ezN69e4mPj8fH\nx8e87dChQyQkJJCTk4O3tzdjxoyxmNO/rKyMxMRE9u7di5ubG6GhoYwcObLWKXuFEEIIcZFDrvg/\n/vhj+vTpU219eXk5CxcuJDw8nMTERKKioli6dCmFhYXmfdatW0deXh7x8fHExcWRmZnJ1q1b7Rnf\nTLX2JNXygmQW1qlYxipmVo2KZXzNtPHPnTuX/v37V1u/e/duNE0jPDwcgHbt2hESEkJKSgpw8YvB\n9u3bGTp0KJqmoWkaERERDqv4hRBCCNU4VRt/VlYWQUFBpKWlMXr0aNLT0wkMDCQrKwuA3NxcSktL\n8ff3Jyoqiri4OAIDAzl58iQmk8nueVVrT1ItL0hmYZ2KZaxiZtWoWMbX/Dj+oqIijEYjBoMBXdcx\nGAxomma+1V9UVASA0WhE13V0XUfTNHRdN28TQgghRM0c0rmvJl5eXmRmZhIcHExSUhIAW7ZswdPT\nE8D8s7KykiVLlgBQUlJisc2eVGtPUi0vSGZhnYplrGJm1ahYxtdMG39NLr2tXyUrK4s2bdoA4Ofn\nh4eHB8eOHbPYHhAQgItL7ady6e2U1NRUWZblq3q5oLCAxlJQWODw85PlK19uDJf+vTn6/GT58r9v\ng67req17NBJd13n88ccthvNVVFQwZcoUhg8fTlhYGBkZGcTExBAbG4uXlxcAq1evJisri5kzZwIQ\nGxtLcHAwgwcPrvG9kpOT6dWrV4OfQ2pqqlLfMFXLC5L5Sh0uzmPl0T2NcuzxbUPo0My3UY5tK2co\n47pyhsyN9XfhDH8T4BxlXFeNlTk9PZ2wsDCr2+x+q//8+fNMnjzZvDxt2jQMBgMffPABbm5uREdH\nk5CQwNq1a/H29iYyMtJc6QOMGDGChIQEIiMjcXV1JTQ0lEGDBtn7NIQQQgglOeyK354a64pfCGd1\ntV/xiytztV/xi/+p7Yrfqdr4hRBCCNG4pOKvh8buMNPQVMsLkllYp2IZq5hZNSqWsSMyS8UvhBBC\nXEOk4q8H1XqPqpYXJLOwTsUyVjGzalQs42t+HL8QQgghGpdU/PWgWnuSanlBMgvrVCxjFTOrRsUy\nljZ+IYQQQjQqqfjrQbX2JNXygmQW1qlYxipmVo2KZSxt/EIIIYRoVEpW/J999hmTJk1i3LhxxMbG\nOuyRvKq1J6mWFySzsE7FMlYxs2pULGNHZHaqx/LaYs+ePfz973/n9ddfp1WrVvz1r39l5cqVTJs2\nrUGOf+bCOc5eKLVpX2P7AA4X59m0bwvtOnw0Y32iCSGEEPWmXMX/9ddfEx4eTqtWrQAYOnQo48eP\nJz8/n+bNm9f7+GcvlNZtLusztu02vm2Iwyt+af+yDxUzq0bFMlYxs2pULGNp47dBVlYWQUFBvPvu\nuzz33HNUVFTg7+/P8ePHHR1NCCGEcHrKVfxFRUUYjRevnHVdx2AwoGkahYWFDk7m/KT9yz5UzKwa\nFctYxcyqUbGMHZFZucfyTpo0iWeeeYaQkBDzuujoaJ566im6d+9u9TXJycn2iieEEEI4hZoey6tc\nG39gYCCZmZnmiv/ChQvk5ORwww031Piamk5eCCGEuNYod6t/0KBBJCcnk52djclkYvPmzfTo0aNB\nOvYJIYQQVzvlbvUDfP7552zZsoXKykq6dOnCpEmT8PT0dHQsIYQQwukpWfELIYQQ4sood6tfCCGE\nEFdOKn4hhBDiGiIVfz2YTCby8mybsldcO0wmEwUFBY6OcdUpLi7m6NGjjo5x1Tp79ixffvkla9as\nAaCiooLMzEwHp6q7iooKR0eoVUVFBXl5eeTm5lr8syflhvM5yq+//kpCQgK///47JpPJvN7FxYV1\n69Y5MJl1+/fvp2vXrri5uZGdnU18fDyapjFp0iRatGjh6Hg1Onv2LDt37iQ/P59Ro0ZRUVHBiRMn\nCAoKcnQ0qz744AOeeeYZi3UFBQW89957vPbaa44JVYsdO3Zw++23o2kaGRkZvP3222iaxvTp02sd\nEutIxcXFxMfH89NPP+Hu7k5iYiIlJSWsWLGiwZ7R0VCeffZZ838bDAb+2IXKYDDw3nvv2TvWZf38\n88+899579O7dm7S0NEaNGoXJZOL9998nNjbW0fGs2rdvH7fccovFupKSEhYtWuSU/+8BbNmyhbVr\n11JRUYGLi+V1tz3rEbnit9HKlSvp27cvixYtomvXrixdupS+ffsya9YsR0ezKj4+HoPBAFzM3rVr\nV9q0acOHH37o4GQ1+/nnn5k1axYnTpxgx44dAOYPH2f1/fffV1vXokULcnJyHJDm8tavX4+7uzsA\nH374IY8++ij33XcfSUlJDk5Ws7/85S+0a9eOVatW4eHhAUDTpk05fPiwg5NVN3HiRCZOnEj//v0J\nCAhg0qRJTJw4kQkTJuDp6ck999zj6IhWrVq1ildeeYUJEyaY/z40TePcuXMOTlazxMREMjIyzMu5\nubnMnj2bgIAAB6aq3caNG5k9ezbr1q2r9s+e5IrfRsXFxQwdOhS4WBn5+fkxbtw43njjjWrfOp2B\nyWTC1dWVkpISjh8/zksvvYTJZGLq1KmOjlajqg+foKAgfvzxR8B5P3z++9//ous6FRUV/Pe//7XY\nduTIEae93Vg1zXVBQQH5+fmEhYWh6zqbNm1ydLQa/fbbb1av7C+98+YsbrrpJgDWrFnDq6++ynXX\nXWfe9vzzz7N48WIeeOABR8WrUXFxcbW7auXl5dWuSp1JVXnOnj2bc+fOERMTw+DBg3nkkUccHa1G\nzZs3p2vXro6OIRW/rTw9PamoqMDNzQ03Nzdyc3Px8fGxe9uMrdzd3cnPz2f//v3cfPPNGAwGXF1d\nHR2rVip9+FTdhSguLq52R6Jp06Y8/fTTjoh1Wddddx2nTp3ixx9/pFevXgDmO0POymg0cvToUdq2\nbWted+DAAfMTOp1Rbm4umqZZrPPx8eHkyZMOSlS7Nm3a8MknnzBs2DDzur/+9a/06NHDgalq1759\ne55++mneeOMNzp07x9NPP80dd9zh6Fi1GjhwINu2bSM8PNyhOaTit1GvXr346aefuPXWW+nXrx9v\nvPEGXl5edOjQwdHRrBowYAAvv/wy586d4+WXXwYutp9XPeDIGan04bN06VIAJk+ebP5vFdx///1E\nR0fj6urK/PnzAcjOzsbLy8vByWo2bNgw5s6dy4ABAzh//jwbN25k69atTte+f6lOnToRFxfHE088\ngZ+fH9nZ2fz1r3+1+PLiTMaNG0dMTAxbt27l/PnzPPfcczRv3pyZM2c6Olo1Z87871nobdu25a67\n7uKHH36ga9eu5m0+Pj6Oiler9PR0Dh48yK5duyzWGwwGXnnlFbvlkAl8bFReXm5u+wLYs2cPZ8+e\npX///uZ2R2fz448/0qJFC/NVdEVFBTk5OVx//fUOTmZdbm4uMTExnD17lvPnz+Pn52f+8GnWrJmj\n41l16NAhOnbs6OgYdXLixAmaNm2Kt7e3ed25c+ec+kvhb7/9xr/+9S/OnDlD69atuffee522MyJc\nfIpoQkICe/bswWQy4ebmxu23386TTz7ptNOLm0wmfvvtN3MZt2/f3tGRrIqIiLjsPhs2bLBDkrrb\nvn17jdsGDBhgtxxS8QunosqHjxC2qKyspKioCC8vL6dsshLXJqn4hRBCiGuItPHbaPny5dXG5Vpb\nnjhxoiPiVbNs2bIat0VGRtoxydXt7NmzbNq0yWJ+h8rKSs6ePUtcXJyD0/3PkiVLmDp1Kh9++KHT\ndjysybRp0xg4cCD9+vVz6r4If6TanBSqKSwsVOrvAS5+Lo8ZM8bhzWpy76kOfv75Z3x9ffH19aVF\nixbs3bsXTdPw9fXFx8fHqTqUtG7dGn9/f4ufv/32m8XwImezY8cOLly4AEBGRgZTpkzh+eef5/jx\n4w5OVrMlS5aQnZ1Nnz59KC4u5u6776a0tJQhQ4Y4OpqF/fv3Axf7pqjmiSee4ODBg0RFRfHOO++w\nb9++ahPjOBtV5qSo+rv45ptvSE1NtfrPWb344ouUlpY6Okad/Pjjjw6v9EGu+G129uxZXnrpJfz9\n/c3rbr75Zr788stqM7c5g6o5By4VGhrKxo0bHZDGNuvXryc0NBT43+QylZWVJCUl8dJLLzk4nXVH\njx5l+fLluLu7880339CvXz/at2/PRx99xMCBAx0dz8L3339vdd6BKl26dLFzItv07t2b3r17U1JS\nwq5du9i0aRMrV64kNDSUxx9/3NHxrFJlToodO3bQvXt3li9fTufOna3uc9ddd9k5lW1MJpNTX8hY\n06JFC7Kzsx0+FFUqfhtlZGRYVPoAXbt25e2333ZQorpr1aqVUz9bQMXJZby8vCxGe5w/f54bbrjB\n6eY4HzZsGOvXr7c670AVZx+W2LRpU2655RaKior46quv+P7775224ldlToqqZj8vLy/mzJnj4DR1\n07NnT3bs2EG/fv0cHcVmjz/+OHFxcUyYMIGmTZtabLPnHWOp+G3k6+vLli1buP/++83rtm3b5rTz\n3v/6668WE7Pouk5mZibZ2dkOTFU7FSeX6dixo3lI30033cSKFSto1aqV0w0/DAsLIywsTLl5BwBK\nS0v57rvv2LFjB0ePHqVv375MmzaNTp06OTpajVSakwKodlGjAl9fXz766COOHTtWrRJ99NFHHZSq\ndgsWLABgxowZ1bbZcwii9Oq30dGjR1m0aBEmk4lWrVqRm5uLruvMmDHDKYecTZkyxWLZxcUFHx8f\nhg0bRvfu3R2UqnbJycmsXr3aPLnMn/70J7Kzs3n77bdZuHCho+NZlZ2djaZpNG/enAsXLvDxxx9z\n9uxZhg8f7pTjzP/yl78wYcIER8eokyeffJIOHTpw9913c/vtt9OkSRNHR7osFeekUE3VF1hrFwfS\ngbl2UvHXQUVFBQcPHiQ/Px8fHx86duyIm5vcNGlIKk4uIxrXqVOnlLwilTkphLOSil8IIYS4hsjl\n6lXszJkz7Nixg9zcXFq2bEloaKhTDTm8Gpw8eZLWrVtbdNq6cOECp06dcqrx2t9884153oma+k04\na+9tFak2xnzt2rU89thj1R4slJSUxFNPPeWgVFenjIwMq/N+WBuJ1Vik4r8C3333HbfffrujY9Tq\nv//9LzExMdx88834+/tz+PBhNm3axAsvvOAUj4W0RpXJcC711ltvMX/+fIvORRUVFbz33ntONeIj\nOTkZg8FARUUFR44cMXeMM5lMHDx4kC5dukjF34BefPFFYmNjlRlu9sUXX7B3716ioqIsHiS0a9cu\np634Vfy82LBhA1u2bKFNmzacPHmSNm3acPDgQfr06WPXHFLxX4HVq1c7fcWflJTE888/T8+ePc3r\n9u3bx0cffcQbb7zhwGQ1W7JkCe7u7vTp04evv/6aBx98kC+++MLpJsO51Pnz56v1KDYajU43Xvu1\n114DICEhgaFDh3LLLbeYt3399decPn3aQclqtnnzZh555BGWL19udbszzZT5R6qNMW/evDlPP/00\nsbGx/PnPf+bPf/6zoyNdloqfF8nJycTExODv78+8efOYPXs2KSkpnDp1yq45nGtQqWgweXl5FpU+\nwC233EJubq6DEl3e0aNHmT59Ovfeey8eHh7069eP5557jt27dzs6Wo3c3d3Nsw1WOXfunNN2+vzh\nhx8sKn24ONTv22+/dVCimlV16Ny1a5d5dsyqmTOdbabMP6oaY66S7t278+abb/Kf//yHmJgYCgsL\nHR2pVip+XjRp0oSAgABcXFyoqKgAoH///tUe09vYnPPTSdSbp6cnx44do02bNuZ1R44cwdPT04Gp\naqfKZDiX6tWrF6tWrWLcuHG4uLhw4cIFEhISzPMQOBsXFxd+++03izHwx44do7y83IGprLv33nuB\ni18ALh0PrwIVx5jDxf8HZ82axZYtW3jxxRed7s7VpVT8vPDz8yM/P5/mzZvj5+fHd999R5s2bew+\n9bBU/FfA2W/zAwwfPpzXX3+d8PBwWrVqxalTp0hOTnbaW6OgzmQ4lxo+fDhvv/02EydOxN/fnxMn\nTtC2bVvGjRvn6GhWDRs2jHnz5tGnTx9atmxJdnY2aWlpjBo1ytHRalQ1jbNKcnJyzLMMFhUVOTrO\nZV3arm8wGHjggQfo1q0bK1eudFyoy1Dx82Lw4MHk5OTQvHlzHn30UebOnUtJSQljxoyxaw4ZzncV\nO3ToEP/617/Iy8vD19eXgQMH0rFjR0fHqpFqk+FU0XWdI0eOcPr0aSXGa2dlZbFr1y4KCgrw9fXl\n9ttvJyAgwNGxhKgTVT8vLlVZWcmFCxfs3h9EKn4hhBDCDqpu8zuaVPxXqZKSEn7++WdKSkqqPcLU\n2Z4ap7Jvv/2Wjz/+2GqnSXvOvW0r1fKCmpmrnDlzxnxnxdnH9WdkZJCcnGye92PgwIFOf/dKNaNG\njaJ79+4MGDCA3r17O6wTsFT8V6nnnnsOuNjJ6I8TtsyePdsRkS6rsrKSr7/+mj179lBYWIifnx/9\n+vJVtSAAAA+HSURBVPXjjjvucHS0Gk2YMIFRo0bRtWtXXF1dLbb5+vo6KFXNVMsLamY+deoU77//\nvrlDbVFREd27d2fixIn4+fk5Ol41aWlpxMfHExYWZu4T9K9//YtJkyYREhLi6HhWmUwmtm3bxu7d\nuykoKMDPz4+77rrLqeejKC4uZteuXaSmpnL8+HHuvPNO7r77btq1a2fXHNK57zKqZuA6c+ZMjfs4\n47Cili1b8uyzzzr9VcalPvzwQ3755Rfuv/9+fHx8OHXqFOvWrSMjI8NpO59dd911Sj0WVLW8oGbm\nJUuWcNNNNzFnzhyaNGnCuXPnWL9+Pe+//z7z5s1zdLxqNmzYwCuvvGJRAd1xxx0sW7bMaSv+VatW\nsX//fv785z/j5+fH6dOn+eSTT8jIyHDaSYeaNWvGPffcwz333ENeXh7ffvstcXFxuLq6EhYWxoAB\nA/Dw8Gj0HFLxX8bixYuZM2cOkydPrnEfZ7zd+Nhjj/HSSy/RrVu3atuc9clVu3fv5p133rEYchgS\nEsL06dOdtuLv27cv33//Pbfddpujo9hEtbygZubTp0/z+uuvm++2GY1GxowZw/jx4x2czLqioqJq\nV53t27d36hEJu3bt4u2337a4uAkJCWHatGlOW/FXqaio4PDhw/z222+cPXuWnj17cvr0aWbOnMmE\nCRMa/fHNUvFfxqxZs4CLV/Xx8fEOTmO7RYsWcdttt+Hn52dxq9+Zn2/v7e1dbcyzt7e3U3SG+aPl\ny5djMBiorKzkq6++Ij093aIvhbPNKqdaXlAzc5WgoCAKCgos/nbz8vKc6vkNl/Lx8WH//v0Wj+z+\nz3/+47RNKXDxs+GPQ/e8vLyc8vOiyk8//URqaip79uzhT3/6E/3792fixInmz717772X+fPnmx85\n3Fik4r8Me9x2aQx33303t912G126dHF0FJsNHjyYTZs2mR9WYTKZSExMZPjw4Q5OVp2Pj4/5S9Tg\nwYMtvlDV9iAcR1EtL6iZuUrHjh2JjY1l8ODBwMUrvE8//ZSwsDBSU1PN+Z2lPfqpp55i4cKF3Hrr\nreY2/n379pkvfJyRSp8XVZYtW0b//v158803uf7666ttb926tV06/EnnPhup9rSt6dOnc+rUqWrf\n2A0GA++9956DUlk3YsQI83+bTCaLJ91VLa9bt84R0YS4Iq+99lq1LybWvqzMmTPHnrFqlZubyzff\nfENeXp65o5wzdkSs8vjjj6PrutXPi0s502eHtXx/ZI8hf1LxX6V+/vnnGrfddNNNdkxyednZ2Zfd\np1WrVnZIIoRQhS2fGyCfHdZIxS+EEAJd1zl8+DC5ubn4+fnRoUMHp25OEVdO2vhttG3bNvr164em\naY6OYpOEhATuvvtuOnTo4OgoNlN5ohYh/igjI8Pqs+Kr2qSdyZkzZ1i4cCH5+fm0atWK7Oxsmjdv\nTnR0tNN28Nu/fz9du3bFzc2N7Oxs4uPj0TSNSZMm0aJFC0fHc2pS8dvor3/9q1Iz3mmaRmxsLEaj\nkQEDBtCvXz+8vb0dHatWq1atqnGiFiFUsmHDBrZs2UKbNm04efIkbdq04eDBg/Tp08fR0axKTEzk\ntttu49FHH8XFxQWTycTf/vY3EhMTiY6OdnQ8q+Lj43n//fcBWLlyJV27dqWyspIPP/yQadOmOTid\nc5OK30YeHh6Ul5fTpEkTR0exyVNPPcWoUaM4cOAAO3fuZMaMGXTq1IkBAwYQHBzslBWrihO1CGFN\ncnIyMTEx+Pv7M2/ePGbPnk1KSgqnTp1ydDSrMjIymDFjhvnWvouLC4899hhTpkxxcLKamUwmXF1d\nKSkp4fjx47z00kuYTCamTp3q6GhOr/buhcIsIiKC5cuXm2/bqcDFxcU8TejLL79MXl4eb7/9NlFR\nUWzevNnpzqVqohYhVNekSRMCAgJwcXGhoqICgP79+7Nr1y4HJ7PO3d2dkpISi3XFxcUOm0veFu7u\n7uTn5/PDDz9w8803YzAYnPKCxhk572/Vyaxbt46CggImTZpkMbbfGYfHVcnNzSU1NZWdO3dSVFRE\naGgoUVFReHt7m6cP/b//+z9Hx1R6ohbx/9q7t5Ao3gYM4I+lttpmbiseNlGIEtOyIsq0QLPIG6EL\nE6VaT0VlFFhBQWaoROcgsjJQO6gVRgc6GF2k2AEqoyLpIBVGomm562rtmq3LznfR535su//a+v46\no/P87nZmlGdvfJx533lfcsbPz8/2Wpafnx8ePnyI0NBQfPv2TexoTsXHx2PPnj1ITU1FQEAA2tvb\nUV1djbi4OLGj/aP4+Hjk5eWht7cXeXl5AACDwQBvb2+Rk0kfZ/W7aDi9HgcAO3fuxPv37zFnzhzE\nxcVh+vTpDu+P5uTkSGI1wgsXLtgeMf78rvPA55SUFLHiEf2xR48eYcKECZgyZQpaWlpQWFgIk8mE\nzMxM26I+UiIIAm7cuIHa2lro9Xqo1WokJCQgKSnpt++di+n58+dQqVS2FREtFgs6OzudLo5D/8Pi\nH6Fqa2sRGxsLLy8vp+cFQUB1dTXS0tKGOBmR/FitVpjN5mG7EiiNLCx+IqJ/2adPn/D48WOYTCaH\np1ipqakiJiPiGP8fEQQBBoPB4b3c8PBwkZM5Z7FY0NPTg5//t5PyMpxEI8HOnTsRHBzssEkWkRSw\n+F3U0NCAkpIS9Pb2AvgxY14QBKjV6kHfSelv1NTU4Ny5c7BYLJJeu5poJJo2bRpSUlIQGBgodhQi\nByx+F1VWViIrKwvR0dHYv38/8vPzcerUKYSFhYkdzanLly8jPz8fYWFhkp6cQzQSxcTEYPv27QgN\nDXU4J6WNeUieWPwuslgstsVlzGYzAGD58uXYsWMH5s+fL2Y0p3x9fSU7BEE00pWUlCApKQlqtZqP\n+klyWPwu8vHxgdlshqenJxQKBVpbW+Hv74/u7m6xozmVkJCA27dvY/HixWJHIZKd5ORkqFQqxMTE\niB2FyAGL30Xz5s1DU1MToqKikJiYiKKiIigUCsycOVPsaE49ffoUb968cVgpzM3NDTt27BApFZE8\nnD9/Hv39/ba15H8+RyQmvs73l5qbm9Hd3Y2ZM2dKcgy9vr7+H8/Fx8cPWQ4iOfrVXvHcH57ExuIn\nIiKSEendqkqcXq/H+/fv8eXLF7GjuGw4ZSUiosHFMX4Xffz4EcXFxfjw4QOUSiW+fv2K6dOnY82a\nNZJcEKe/vx9VVVWoq6uzTUpcuHAhtFotPDw8xI5HREQi4aN+F+Xl5SEyMhLLli2Dp6cnent7ce7c\nObS0tKCoqEjseA7OnDmD9vZ2pKenw9/fHx0dHaioqIBGo0FmZqbY8YiISCR81O+iz58/Y/ny5fD0\n9AQAeHt7Izs7G21tbSInc+7x48fYtGkTNBoN3N3dERwcjM2bN3O/eyIimWPxuygkJMThnX29Xm/b\nDlJqBEFweKTv4eFh22eAiIjkiWP8v3Hv3j24ublh8uTJOHjwoG0vbYvFgkuXLmHRokUiJ3QuIiIC\nJ0+ehFarxZgxY9DX14eKigpERkaKHY2IiETEMf7fKCgosC25+fP2mgOfpbj2ttFoxNGjR9HY2Ijx\n48ejp6cHUVFR2LBhA5RKpdjxiIhIJCz+Ec5gMECv10OtVkOlUokdh4iIRMbiJyIikhGO8buopqYG\n9fX1MBqNDudKSkpESPRrgiDg9evX0Ol0GPjfbmBoIi4uTuR0REQkFha/i65evQqtVgu1Wi3Jtfl/\nduTIETQ2NiIkJMQhL4ufiEi+WPwuyszMRFdXF6ZOnSp2FJc0NjaiuLgY3t7eYkchIiIJYfG76MmT\nJ7h//z6qqqrsZvYD0txmMyAgAKNHjxY7BhERSQyL30Xv3r3DsWPHJLkuvzNpaWk4e/YssrOzxY5C\nREQSwln9Lrp48SKam5sRGxvrcG7BggUiJPq1tLQ0CILgdD6CFJ9QEBHR0OAdv4tevHgBAKitrbUd\nG5glL8XiP3LkiNgRiIhIgnjH/3+wWq3o6uoaNo//rVYrDAYD1Gq12FGIiEgkvON3UVNTE8rKytDW\n1ma30c2oUaMk+eh8uOUlIqKhweJ3UWlpKWJjYxEdHY3S0lJs3LgRVVVViI+PFzuaU8MtLxERDQ3p\nr0QjEUajEcnJyQgODobVaoWfnx9Wr16N6upqsaM5NdzyEhHR0GDxu2jcuHGwWCwAAHd3d+h0Onh7\ne0On04mczLnhlpeIiIbG6IKCggKxQwwHOp0OVqsVQUFBEAQBZWVlePjwIfz9/SU5q3+45SUioqHB\nWf0u6u/vh4eHh+1zQ0MDDAYD4uLioFAoREzm3HDLS0REQ4PFT0REJCMc4yciIpIRFj8REZGMsPiJ\niIhkhMVPREQkIyx+IiIiGWHxExERyQiLn4jsvHz5EqmpqTAajXbHU1NT8ejRI9vnzs5OFBcXY926\ndVixYgXWr1+PmzdvOvw+k8mE48ePIysrCxkZGTh8+DC+fPnicF19fT3S09PR3d2N3bt3Q6vVIicn\nx7YlNhH9O7hJDxH9lV27dsHPzw8bNmyASqVCV1cXlEql3TVWqxV79+6FUqnEtm3b4O7ujurqahw6\ndAiFhYUOv1MQBBw4cABz587FypUr0dLSgrCwsKH6SkSywOInoj9mNBrR0dGBVatWYdq0aQCAiRMn\nOlz37NkztLa2orS0FO7uP/7cbNmyBatXr0ZTUxPCw8PtrjebzUhISMCiRYsAACEhIYP8TYjkh8VP\nRH9MqVQiKioKJSUlWLJkCWbPnu20pF+9eoXe3l5kZGTYHbdYLGhtbXUofk9PT1vpE9HgYPET0V/Z\nunUr6urq8ODBA1y4cAEajQarVq1CRESE7RqTyYRJkyZh48aNDj/v4+PjcMzNzW1QMxMRi5+I/pKH\nhwcSExORmJgIg8GAiooK7Nu3DydOnICXlxeAH+Wu1+sRFBTEUieSCM7qJyI7A6VtMplsx/R6/S9/\nRqVSYenSpejr67O7dsaMGejp6cHdu3cHJywR/THe8RORHY1GA4VCgTt37iAlJQVv375FeXm5w3U1\nNTUIDQ2Fr68vdDodLl26hMDAQGg0Gts1kZGRmDNnDsrLy/H9+3dERETg69evePPmDZYuXTqUX4uI\n/ovFT0R2FAoF1q5di8rKSty6dQsRERHIzc1Fbm6u3XVv377FtWvXYDQa4ePjg1mzZiElJQWjRtk/\nSMzNzcWVK1dw/fp1nD59GmPHjkVYWBj6+vqgUCjsruVwANHgcxMEQRA7BBEREQ0NjvETERHJCIuf\niIhIRlj8REREMsLiJyIikhEWPxERkYyw+ImIiGSExU9ERCQjLH4iIiIZYfETERHJyH8AKTLjqt3m\nEpQAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x109bc6198>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt_size_user = df['size_mb'].groupby(df['user']).agg(sum)\n",
    "plt.xlabel('User')\n",
    "plt.ylabel('Size (MB)')\n",
    "plt.title('Size per User')\n",
    "plt_size_user.plot(kind='bar', figsize=(8,3))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['app-logs', 'apps', 'hdp', 'mapred', 'mr-history', 'ranger', 'tmp',\n",
       "       'user'], dtype=object)"
      ]
     },
     "execution_count": 101,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['path_main'] = df.fname.apply(lambda fname: fname[1:fname.find('/', 1)] if fname.find('/', 1) > 0 else fname[1:])\n",
    "df['path_main'].unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x10a92beb8>"
      ]
     },
     "execution_count": 102,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAf4AAAE1CAYAAAAGQIJSAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10a329c50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.xlabel('Main Path')\n",
    "plt.ylabel('Size (MB)')\n",
    "plt.title('Size per Main Path')\n",
    "df.groupby(['path_main'])['size_mb'].sum().plot(kind='bar', figsize=(8,3))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>permissions</th>\n",
       "      <th>replicas</th>\n",
       "      <th>user</th>\n",
       "      <th>group</th>\n",
       "      <th>size</th>\n",
       "      <th>mod_date</th>\n",
       "      <th>mod_time</th>\n",
       "      <th>fname</th>\n",
       "      <th>src</th>\n",
       "      <th>is_dir</th>\n",
       "      <th>num_replicas</th>\n",
       "      <th>dt_mod</th>\n",
       "      <th>size_mb</th>\n",
       "      <th>path_main</th>\n",
       "      <th>fname_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>drwxrwxrwx</td>\n",
       "      <td>-</td>\n",
       "      <td>yarn</td>\n",
       "      <td>hadoop</td>\n",
       "      <td>0</td>\n",
       "      <td>2015-10-20</td>\n",
       "      <td>13:26</td>\n",
       "      <td>/app-logs</td>\n",
       "      <td>20151024</td>\n",
       "      <td>True</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2015-10-20 13:26:00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>app-logs</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>drwxrwx---</td>\n",
       "      <td>-</td>\n",
       "      <td>ambari-qa</td>\n",
       "      <td>hadoop</td>\n",
       "      <td>0</td>\n",
       "      <td>2015-10-20</td>\n",
       "      <td>13:26</td>\n",
       "      <td>/app-logs/ambari-qa</td>\n",
       "      <td>20151024</td>\n",
       "      <td>True</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2015-10-20 13:26:00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>app-logs</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>drwxrwx---</td>\n",
       "      <td>-</td>\n",
       "      <td>ambari-qa</td>\n",
       "      <td>hadoop</td>\n",
       "      <td>0</td>\n",
       "      <td>2015-10-21</td>\n",
       "      <td>00:47</td>\n",
       "      <td>/app-logs/ambari-qa/logs</td>\n",
       "      <td>20151024</td>\n",
       "      <td>True</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2015-10-21 00:47:00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>app-logs</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>drwxrwx---</td>\n",
       "      <td>-</td>\n",
       "      <td>ambari-qa</td>\n",
       "      <td>hadoop</td>\n",
       "      <td>0</td>\n",
       "      <td>2015-10-20</td>\n",
       "      <td>13:26</td>\n",
       "      <td>/app-logs/ambari-qa/logs/application_144537273...</td>\n",
       "      <td>20151024</td>\n",
       "      <td>True</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2015-10-20 13:26:00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>app-logs</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>-rw-r-----</td>\n",
       "      <td>2</td>\n",
       "      <td>ambari-qa</td>\n",
       "      <td>hadoop</td>\n",
       "      <td>384</td>\n",
       "      <td>2015-10-20</td>\n",
       "      <td>13:26</td>\n",
       "      <td>/app-logs/ambari-qa/logs/application_144537273...</td>\n",
       "      <td>20151024</td>\n",
       "      <td>False</td>\n",
       "      <td>2</td>\n",
       "      <td>2015-10-20 13:26:00</td>\n",
       "      <td>0.000366</td>\n",
       "      <td>app-logs</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  permissions replicas       user   group  size    mod_date mod_time  \\\n",
       "0  drwxrwxrwx        -       yarn  hadoop     0  2015-10-20    13:26   \n",
       "1  drwxrwx---        -  ambari-qa  hadoop     0  2015-10-20    13:26   \n",
       "2  drwxrwx---        -  ambari-qa  hadoop     0  2015-10-21    00:47   \n",
       "3  drwxrwx---        -  ambari-qa  hadoop     0  2015-10-20    13:26   \n",
       "4  -rw-r-----        2  ambari-qa  hadoop   384  2015-10-20    13:26   \n",
       "\n",
       "                                               fname       src is_dir  \\\n",
       "0                                          /app-logs  20151024   True   \n",
       "1                                /app-logs/ambari-qa  20151024   True   \n",
       "2                           /app-logs/ambari-qa/logs  20151024   True   \n",
       "3  /app-logs/ambari-qa/logs/application_144537273...  20151024   True   \n",
       "4  /app-logs/ambari-qa/logs/application_144537273...  20151024  False   \n",
       "\n",
       "   num_replicas              dt_mod   size_mb path_main fname_id  \n",
       "0           NaN 2015-10-20 13:26:00  0.000000  app-logs      NaN  \n",
       "1           NaN 2015-10-20 13:26:00  0.000000  app-logs      NaN  \n",
       "2           NaN 2015-10-21 00:47:00  0.000000  app-logs      NaN  \n",
       "3           NaN 2015-10-20 13:26:00  0.000000  app-logs      NaN  \n",
       "4             2 2015-10-20 13:26:00  0.000366  app-logs      NaN  "
      ]
     },
     "execution_count": 103,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# calculates something like the user or group based on the prefix of a path\n",
    "# E.g.: All Projects all located in /projects/ => /projects/proj_a, /projects/proj_b, etc.\n",
    "# Fname_id = proj_a, proj_b,... => size,# files etc. can be calculated based on this id\n",
    "df['fname_id'] = \"\"\n",
    "search = '/user/'\n",
    "selection = df['fname'].map(lambda x: x.startswith(search))\n",
    "df['fname_id'] = df[selection]['fname'].map(lambda x: x[len(search): x.find('/',len(search))+1].replace('/',''))\n",
    "\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "fname_id\n",
       "               0.000000\n",
       "ambari-qa      0.526881\n",
       "oozie        602.341820\n",
       "Name: size_mb, dtype: float64"
      ]
     },
     "execution_count": 104,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# quota per filename id\n",
    "df.groupby(['fname_id'])['size_mb'].sum()"
   ]
  }
 ],
 "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": "3.4.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}