From 361c5a97f2c5784972a91106bf6076d62642cbd6 Mon Sep 17 00:00:00 2001
From: frankknoll <Knoll_Frank@web.de>
Date: Mon, 1 May 2023 15:00:26 +0200
Subject: [PATCH] getCountryCountsByBatchcodeTable()

---
 src/CountriesByBatchcodeProvider.py |   65 +-
 src/HowBadIsMyBatch.ipynb           | 1051 +--------------------------
 2 files changed, 44 insertions(+), 1072 deletions(-)

diff --git a/src/CountriesByBatchcodeProvider.py b/src/CountriesByBatchcodeProvider.py
index 4b41282348b..81e0f26a0cd 100644
--- a/src/CountriesByBatchcodeProvider.py
+++ b/src/CountriesByBatchcodeProvider.py
@@ -1,39 +1,50 @@
 import pandas as pd
-from CompletedBatchcodeColumnAdder import CompletedBatchcodeColumnAdder
-from BatchcodeCompletion import BatchcodeCompletion
-from CountriesColumnAdder import CountriesColumnAdder
 from BatchCodeTableFactory import BatchCodeTableFactory
 from InternationalVaersCovid19Provider import getInternationalVaersCovid19
 
-def getCountriesByCompletedBatchcode(internationalVaersCovid19):
-    result = _readExploration('data/Country By Batchcode Search Term.csv', indexName = 'Batchcode Search Term')
-    result = _addCompletedBatchcodeColumn(result, internationalVaersCovid19)
-    columnName = 'Countries'
-    result = CountriesColumnAdder().addCountriesColumn(result, columnName = columnName)
-    return result[[columnName]].droplevel('Batchcode Search Term')
 
-def _addCompletedBatchcodeColumn(country_By_Batchcode_Search_Term, internationalVaersCovid19):
-    return CompletedBatchcodeColumnAdder(_getCompleteBatchcode(internationalVaersCovid19)).addCompletedBatchcodeColumn(country_By_Batchcode_Search_Term)
+def getCountryCountsByBatchcodeTable():
+    return _combineCountryCountsByBatchcodeTables(
+        countryCountsByClickedBatchcode = _getCountryCountsByClickedBatchcode(),
+        countryCountsByBatchcodeBeforeDeletion = _getCountryCountsByBatchcodeBeforeDeletion())
 
-def _getCompleteBatchcode(internationalVaersCovid19):
-    batchCodeTable = BatchCodeTableFactory(internationalVaersCovid19).createGlobalBatchCodeTable()
-    return BatchcodeCompletion(ADR_by_Batchcode = batchCodeTable).completeBatchcode
-     
-def getCountriesByClickedBatchcode():
-    result = _readExploration('data/Country By Clicked Batchcode.csv', indexName = 'Clicked Batchcode')
-    columnName = 'Countries'
-    result = CountriesColumnAdder().addCountriesColumn(result, columnName = columnName)
-    return result[[columnName]]
 
-def _readExploration(csvFile, indexName):
-    exploration = pd.read_csv(csvFile, header = [0], index_col = 0, skiprows = 6, on_bad_lines = 'warn')
-    exploration.drop(index=indexName, inplace=True)
-    exploration.index.rename(indexName, inplace=True)
-    exploration.drop(columns='Totals', inplace=True)
-    for column in exploration.columns:
-        exploration[column] = exploration[column].astype('int64')
+def _getCountryCountsByClickedBatchcode():
+    exploration = pd.read_csv('data/Country By Clicked Batchcode.csv', index_col = 0, skiprows = [0, 1, 2, 3, 4, 5, 7])
+    exploration.index.name = 'VAX_LOT'
+    exploration.rename(
+        columns =
+        {
+            'Country': 'COUNTRY',
+            'Event count': 'COUNTRY_COUNT_BY_VAX_LOT'
+        },
+        inplace = True)
+    exploration.set_index('COUNTRY', append = True, inplace = True)
     return exploration
 
+
+def _getCountryCountsByBatchcodeBeforeDeletion():
+    internationalVaersCovid19 = getInternationalVaersCovid19(dataDir = 'VAERS/VAERSBeforeDeletion', years = [2020, 2021, 2022])
+    return (internationalVaersCovid19
+            .groupby('VAX_LOT')
+            ['COUNTRY'].value_counts()
+            .to_frame(name = 'COUNTRY_COUNT_BY_VAX_LOT'))
+
+
+def _combineCountryCountsByBatchcodeTables(countryCountsByClickedBatchcode, countryCountsByBatchcodeBeforeDeletion):
+    countryCountsByBatchcode = pd.merge(
+        countryCountsByClickedBatchcode,
+        countryCountsByBatchcodeBeforeDeletion,
+        how = 'outer',
+        left_index = True,
+        right_index = True,
+        suffixes=(' Clicked', ' Before Deletion'))
+    countryCountsByBatchcode.fillna(0, inplace = True)
+    for column in countryCountsByBatchcode.columns:
+        countryCountsByBatchcode[column] = countryCountsByBatchcode[column].astype('int64')
+    return countryCountsByBatchcode
+
+
 def getCountriesByBatchcodeBeforeDeletion():
     internationalVaersCovid19 = getInternationalVaersCovid19(dataDir = 'VAERS/VAERSBeforeDeletion', years = [2020, 2021, 2022])
     batchCodeTable = BatchCodeTableFactory(internationalVaersCovid19).createGlobalBatchCodeTable(countriesAsList = True)
diff --git a/src/HowBadIsMyBatch.ipynb b/src/HowBadIsMyBatch.ipynb
index 0c5244e27df..ea00af3157c 100644
--- a/src/HowBadIsMyBatch.ipynb
+++ b/src/HowBadIsMyBatch.ipynb
@@ -15,14 +15,14 @@
     "import os\n",
     "from VAERSFileDownloader import updateVAERSFiles\n",
     "from datetime import datetime\n",
-    "from DateProvider import DateProvider\n",
     "from InternationalVaersCovid19Provider import getInternationalVaersCovid19, get_international_VAERSVAX_VAERSSYMPTOMS_Covid19\n",
     "from BatchCodeTableHtmlUpdater import updateBatchCodeTableHtmlFile\n",
     "from BatchCodeTablePersister import createAndSaveBatchCodeTables\n",
     "from SymptomByBatchcodeTableFactory import SymptomByBatchcodeTableFactory\n",
     "from HistogramFactoryAndPersister import createAndSaveGlobalHistograms, createAndSaveHistogramsForCountries\n",
     "from BatchCodeTableFactory import BatchCodeTableFactory\n",
-    "from BatchCodeTableHavingGuessedCountriesFactory import BatchCodeTableHavingGuessedCountriesFactory\n"
+    "from BatchCodeTableHavingGuessedCountriesFactory import BatchCodeTableHavingGuessedCountriesFactory\n",
+    "from CountriesByBatchcodeProvider import getCountryCountsByBatchcodeTable\n"
    ]
   },
   {
@@ -35,20 +35,6 @@
     "print(datetime.now().strftime(\"%d.%m.%Y, %H:%M:%S Uhr\"))"
    ]
   },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "id": "ffad1c04",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "dateProvider = DateProvider()\n",
-    "print('          lastUpdated:', dateProvider.getLastUpdated())\n",
-    "print('lastUpdatedDataSource:', dateProvider.getLastUpdatedDataSource())\n",
-    "needsUpdate = dateProvider.needsUpdate()\n",
-    "print('needsUpdate:', needsUpdate)"
-   ]
-  },
   {
    "cell_type": "code",
    "execution_count": null,
@@ -169,1038 +155,13 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 6,
-   "id": "dc8fa52b",
+   "execution_count": null,
+   "id": "d6c2b650",
    "metadata": {},
    "outputs": [],
    "source": [
-    "import pandas as pd\n",
-    "from CountriesByBatchcodeProvider import getCountriesByClickedBatchcode, getCountriesByBatchcodeBeforeDeletion"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 2,
-   "id": "b01a5f94",
-   "metadata": {},
-   "outputs": [
-    {
-     "name": "stderr",
-     "output_type": "stream",
-     "text": [
-      "b'Skipping line 9: expected 22 fields, saw 23\\n'\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>Countries</th>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>Clicked Batchcode</th>\n",
-       "      <th></th>\n",
-       "    </tr>\n",
-       "  </thead>\n",
-       "  <tbody>\n",
-       "    <tr>\n",
-       "      <th>(not set)</th>\n",
-       "      <td>{Germany, Poland, United Kingdom, Australia, B...</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>FE6208</th>\n",
-       "      <td>{Germany, Poland, United Kingdom, Australia, B...</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>039K20A</th>\n",
-       "      <td>{Sweden, United States, Spain, France, Austral...</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>EN6201</th>\n",
-       "      <td>{Sweden, United States, Spain, France, Austral...</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>FD6840</th>\n",
-       "      <td>{Sweden, United States, Spain, France, Austral...</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>...</th>\n",
-       "      <td>...</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>ZTC540693</th>\n",
-       "      <td>{United States}</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>ZU5A21A</th>\n",
-       "      <td>{}</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>ZZ3249</th>\n",
-       "      <td>{Hungary}</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>026L20A</th>\n",
-       "      <td>{United States}</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>ÂK4244</th>\n",
-       "      <td>{Hungary}</td>\n",
-       "    </tr>\n",
-       "  </tbody>\n",
-       "</table>\n",
-       "<p>5392 rows × 1 columns</p>\n",
-       "</div>"
-      ],
-      "text/plain": [
-       "                                                           Countries\n",
-       "Clicked Batchcode                                                   \n",
-       "(not set)          {Germany, Poland, United Kingdom, Australia, B...\n",
-       "FE6208             {Germany, Poland, United Kingdom, Australia, B...\n",
-       "039K20A            {Sweden, United States, Spain, France, Austral...\n",
-       "EN6201             {Sweden, United States, Spain, France, Austral...\n",
-       "FD6840             {Sweden, United States, Spain, France, Austral...\n",
-       "...                                                              ...\n",
-       "ZTC540693                                            {United States}\n",
-       "ZU5A21A                                                           {}\n",
-       "ZZ3249                                                     {Hungary}\n",
-       "  026L20A                                            {United States}\n",
-       "ÂK4244                                                     {Hungary}\n",
-       "\n",
-       "[5392 rows x 1 columns]"
-      ]
-     },
-     "execution_count": 2,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
-   "source": [
-    "countriesByClickedBatchcode = getCountriesByClickedBatchcode()\n",
-    "countriesByClickedBatchcode"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 3,
-   "id": "521ee957",
-   "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>Countries</th>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>VAX_LOT</th>\n",
-       "      <th></th>\n",
-       "    </tr>\n",
-       "  </thead>\n",
-       "  <tbody>\n",
-       "    <tr>\n",
-       "      <th>6362737</th>\n",
-       "      <td>[United States]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>SCH2</th>\n",
-       "      <td>[United States]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>NO72A</th>\n",
-       "      <td>[Philippines]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>EW096</th>\n",
-       "      <td>[United States]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>FS1926</th>\n",
-       "      <td>[United States]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>...</th>\n",
-       "      <td>...</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>211221A</th>\n",
-       "      <td>[United States]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>2112D21A</th>\n",
-       "      <td>[United States]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>2114214</th>\n",
-       "      <td>[United States]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>211421A</th>\n",
-       "      <td>[United States]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>Ø94F21A</th>\n",
-       "      <td>[Norway]</td>\n",
-       "    </tr>\n",
-       "  </tbody>\n",
-       "</table>\n",
-       "<p>47901 rows × 1 columns</p>\n",
-       "</div>"
-      ],
-      "text/plain": [
-       "                Countries\n",
-       "VAX_LOT                  \n",
-       "6362737   [United States]\n",
-       "SCH2      [United States]\n",
-       "NO72A       [Philippines]\n",
-       "EW096     [United States]\n",
-       "FS1926    [United States]\n",
-       "...                   ...\n",
-       "211221A   [United States]\n",
-       "2112D21A  [United States]\n",
-       "2114214   [United States]\n",
-       "211421A   [United States]\n",
-       "Ø94F21A          [Norway]\n",
-       "\n",
-       "[47901 rows x 1 columns]"
-      ]
-     },
-     "execution_count": 3,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
-   "source": [
-    "countriesByBatchcodeBeforeDeletion = getCountriesByBatchcodeBeforeDeletion()\n",
-    "countriesByBatchcodeBeforeDeletion"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 41,
-   "id": "c91d1b4d",
-   "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>Countries Clicked</th>\n",
-       "      <th>Countries Before Deletion</th>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>VAX_LOT</th>\n",
-       "      <th></th>\n",
-       "      <th></th>\n",
-       "    </tr>\n",
-       "  </thead>\n",
-       "  <tbody>\n",
-       "    <tr>\n",
-       "      <th>FE6208</th>\n",
-       "      <td>{Germany, Poland, United Kingdom, Australia, B...</td>\n",
-       "      <td>[Austria, Croatia, Czechia, Germany, Greece, H...</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>039K20A</th>\n",
-       "      <td>{Sweden, United States, Spain, France, Austral...</td>\n",
-       "      <td>[United States]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>EN6201</th>\n",
-       "      <td>{Sweden, United States, Spain, France, Austral...</td>\n",
-       "      <td>[United Kingdom, United States]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>FD6840</th>\n",
-       "      <td>{Sweden, United States, Spain, France, Austral...</td>\n",
-       "      <td>[Austria, Croatia, Cyprus, Czechia, Germany, G...</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>FF0680</th>\n",
-       "      <td>{Sweden, United States, Canada, Brazil, German...</td>\n",
-       "      <td>[Austria, Croatia, Czechia, Finland, France, G...</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>...</th>\n",
-       "      <td>...</td>\n",
-       "      <td>...</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>ZTC540693</th>\n",
-       "      <td>{United States}</td>\n",
-       "      <td>[United States]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>ZU5A21A</th>\n",
-       "      <td>{}</td>\n",
-       "      <td>[United States]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>ZZ3249</th>\n",
-       "      <td>{Hungary}</td>\n",
-       "      <td>[United States]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>026L20A</th>\n",
-       "      <td>{United States}</td>\n",
-       "      <td>[United States]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>ÂK4244</th>\n",
-       "      <td>{Hungary}</td>\n",
-       "      <td>[United Kingdom]</td>\n",
-       "    </tr>\n",
-       "  </tbody>\n",
-       "</table>\n",
-       "<p>5188 rows × 2 columns</p>\n",
-       "</div>"
-      ],
-      "text/plain": [
-       "                                           Countries Clicked  \\\n",
-       "VAX_LOT                                                        \n",
-       "FE6208     {Germany, Poland, United Kingdom, Australia, B...   \n",
-       "039K20A    {Sweden, United States, Spain, France, Austral...   \n",
-       "EN6201     {Sweden, United States, Spain, France, Austral...   \n",
-       "FD6840     {Sweden, United States, Spain, France, Austral...   \n",
-       "FF0680     {Sweden, United States, Canada, Brazil, German...   \n",
-       "...                                                      ...   \n",
-       "ZTC540693                                    {United States}   \n",
-       "ZU5A21A                                                   {}   \n",
-       "ZZ3249                                             {Hungary}   \n",
-       "  026L20A                                    {United States}   \n",
-       "ÂK4244                                             {Hungary}   \n",
-       "\n",
-       "                                   Countries Before Deletion  \n",
-       "VAX_LOT                                                       \n",
-       "FE6208     [Austria, Croatia, Czechia, Germany, Greece, H...  \n",
-       "039K20A                                      [United States]  \n",
-       "EN6201                       [United Kingdom, United States]  \n",
-       "FD6840     [Austria, Croatia, Cyprus, Czechia, Germany, G...  \n",
-       "FF0680     [Austria, Croatia, Czechia, Finland, France, G...  \n",
-       "...                                                      ...  \n",
-       "ZTC540693                                    [United States]  \n",
-       "ZU5A21A                                      [United States]  \n",
-       "ZZ3249                                       [United States]  \n",
-       "  026L20A                                    [United States]  \n",
-       "ÂK4244                                      [United Kingdom]  \n",
-       "\n",
-       "[5188 rows x 2 columns]"
-      ]
-     },
-     "execution_count": 41,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
-   "source": [
-    "res = pd.merge(\n",
-    "    countriesByClickedBatchcode,\n",
-    "    countriesByBatchcodeBeforeDeletion,\n",
-    "    how = 'inner',\n",
-    "    left_index = True,\n",
-    "    right_index = True,\n",
-    "    suffixes=(' Clicked', ' Before Deletion'))\n",
-    "res.index.name = 'VAX_LOT'\n",
-    "res"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 42,
-   "id": "df6e5ef1",
-   "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>Countries Clicked</th>\n",
-       "      <th>Countries Before Deletion</th>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>VAX_LOT</th>\n",
-       "      <th></th>\n",
-       "      <th></th>\n",
-       "    </tr>\n",
-       "  </thead>\n",
-       "  <tbody>\n",
-       "    <tr>\n",
-       "      <th>FE6208</th>\n",
-       "      <td>{Germany, Poland, United Kingdom, Australia, B...</td>\n",
-       "      <td>{Lithuania, Spain, Czechia, Greece, Germany, A...</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>039K20A</th>\n",
-       "      <td>{Sweden, United States, Spain, France, Austral...</td>\n",
-       "      <td>{United States}</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>EN6201</th>\n",
-       "      <td>{Sweden, United States, Spain, France, Austral...</td>\n",
-       "      <td>{United Kingdom, United States}</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>FD6840</th>\n",
-       "      <td>{Sweden, United States, Spain, France, Austral...</td>\n",
-       "      <td>{Lithuania, Cyprus, Czechia, Greece, Germany, ...</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>FF0680</th>\n",
-       "      <td>{Sweden, United States, Canada, Brazil, German...</td>\n",
-       "      <td>{Portugal, Sweden, Czechia, France, Finland, G...</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>...</th>\n",
-       "      <td>...</td>\n",
-       "      <td>...</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>ZTC540693</th>\n",
-       "      <td>{United States}</td>\n",
-       "      <td>{United States}</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>ZU5A21A</th>\n",
-       "      <td>{}</td>\n",
-       "      <td>{United States}</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>ZZ3249</th>\n",
-       "      <td>{Hungary}</td>\n",
-       "      <td>{United States}</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>026L20A</th>\n",
-       "      <td>{United States}</td>\n",
-       "      <td>{United States}</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>ÂK4244</th>\n",
-       "      <td>{Hungary}</td>\n",
-       "      <td>{United Kingdom}</td>\n",
-       "    </tr>\n",
-       "  </tbody>\n",
-       "</table>\n",
-       "<p>5188 rows × 2 columns</p>\n",
-       "</div>"
-      ],
-      "text/plain": [
-       "                                           Countries Clicked  \\\n",
-       "VAX_LOT                                                        \n",
-       "FE6208     {Germany, Poland, United Kingdom, Australia, B...   \n",
-       "039K20A    {Sweden, United States, Spain, France, Austral...   \n",
-       "EN6201     {Sweden, United States, Spain, France, Austral...   \n",
-       "FD6840     {Sweden, United States, Spain, France, Austral...   \n",
-       "FF0680     {Sweden, United States, Canada, Brazil, German...   \n",
-       "...                                                      ...   \n",
-       "ZTC540693                                    {United States}   \n",
-       "ZU5A21A                                                   {}   \n",
-       "ZZ3249                                             {Hungary}   \n",
-       "  026L20A                                    {United States}   \n",
-       "ÂK4244                                             {Hungary}   \n",
-       "\n",
-       "                                   Countries Before Deletion  \n",
-       "VAX_LOT                                                       \n",
-       "FE6208     {Lithuania, Spain, Czechia, Greece, Germany, A...  \n",
-       "039K20A                                      {United States}  \n",
-       "EN6201                       {United Kingdom, United States}  \n",
-       "FD6840     {Lithuania, Cyprus, Czechia, Greece, Germany, ...  \n",
-       "FF0680     {Portugal, Sweden, Czechia, France, Finland, G...  \n",
-       "...                                                      ...  \n",
-       "ZTC540693                                    {United States}  \n",
-       "ZU5A21A                                      {United States}  \n",
-       "ZZ3249                                       {United States}  \n",
-       "  026L20A                                    {United States}  \n",
-       "ÂK4244                                      {United Kingdom}  \n",
-       "\n",
-       "[5188 rows x 2 columns]"
-      ]
-     },
-     "execution_count": 42,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
-   "source": [
-    "res['Countries Before Deletion'] = res['Countries Before Deletion'].map(set)\n",
-    "res"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 43,
-   "id": "62d68836",
-   "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>Countries Clicked</th>\n",
-       "      <th>Countries Before Deletion</th>\n",
-       "      <th>Countries Intersection</th>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>VAX_LOT</th>\n",
-       "      <th></th>\n",
-       "      <th></th>\n",
-       "      <th></th>\n",
-       "    </tr>\n",
-       "  </thead>\n",
-       "  <tbody>\n",
-       "    <tr>\n",
-       "      <th>FE6208</th>\n",
-       "      <td>{Germany, Poland, United Kingdom, Australia, B...</td>\n",
-       "      <td>{Lithuania, Spain, Czechia, Greece, Germany, A...</td>\n",
-       "      <td>[Austria, Germany, Hungary, Spain, Sweden, Uni...</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>039K20A</th>\n",
-       "      <td>{Sweden, United States, Spain, France, Austral...</td>\n",
-       "      <td>{United States}</td>\n",
-       "      <td>[United States]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>EN6201</th>\n",
-       "      <td>{Sweden, United States, Spain, France, Austral...</td>\n",
-       "      <td>{United Kingdom, United States}</td>\n",
-       "      <td>[United Kingdom, United States]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>FD6840</th>\n",
-       "      <td>{Sweden, United States, Spain, France, Austral...</td>\n",
-       "      <td>{Lithuania, Cyprus, Czechia, Greece, Germany, ...</td>\n",
-       "      <td>[Austria, Germany, Hungary, Sweden, United Kin...</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>FF0680</th>\n",
-       "      <td>{Sweden, United States, Canada, Brazil, German...</td>\n",
-       "      <td>{Portugal, Sweden, Czechia, France, Finland, G...</td>\n",
-       "      <td>[Austria, Germany, Hungary, Italy, Netherlands...</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>...</th>\n",
-       "      <td>...</td>\n",
-       "      <td>...</td>\n",
-       "      <td>...</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>ZTC540693</th>\n",
-       "      <td>{United States}</td>\n",
-       "      <td>{United States}</td>\n",
-       "      <td>[United States]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>ZU5A21A</th>\n",
-       "      <td>{}</td>\n",
-       "      <td>{United States}</td>\n",
-       "      <td>[]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>ZZ3249</th>\n",
-       "      <td>{Hungary}</td>\n",
-       "      <td>{United States}</td>\n",
-       "      <td>[]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>026L20A</th>\n",
-       "      <td>{United States}</td>\n",
-       "      <td>{United States}</td>\n",
-       "      <td>[United States]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>ÂK4244</th>\n",
-       "      <td>{Hungary}</td>\n",
-       "      <td>{United Kingdom}</td>\n",
-       "      <td>[]</td>\n",
-       "    </tr>\n",
-       "  </tbody>\n",
-       "</table>\n",
-       "<p>5188 rows × 3 columns</p>\n",
-       "</div>"
-      ],
-      "text/plain": [
-       "                                           Countries Clicked  \\\n",
-       "VAX_LOT                                                        \n",
-       "FE6208     {Germany, Poland, United Kingdom, Australia, B...   \n",
-       "039K20A    {Sweden, United States, Spain, France, Austral...   \n",
-       "EN6201     {Sweden, United States, Spain, France, Austral...   \n",
-       "FD6840     {Sweden, United States, Spain, France, Austral...   \n",
-       "FF0680     {Sweden, United States, Canada, Brazil, German...   \n",
-       "...                                                      ...   \n",
-       "ZTC540693                                    {United States}   \n",
-       "ZU5A21A                                                   {}   \n",
-       "ZZ3249                                             {Hungary}   \n",
-       "  026L20A                                    {United States}   \n",
-       "ÂK4244                                             {Hungary}   \n",
-       "\n",
-       "                                   Countries Before Deletion  \\\n",
-       "VAX_LOT                                                        \n",
-       "FE6208     {Lithuania, Spain, Czechia, Greece, Germany, A...   \n",
-       "039K20A                                      {United States}   \n",
-       "EN6201                       {United Kingdom, United States}   \n",
-       "FD6840     {Lithuania, Cyprus, Czechia, Greece, Germany, ...   \n",
-       "FF0680     {Portugal, Sweden, Czechia, France, Finland, G...   \n",
-       "...                                                      ...   \n",
-       "ZTC540693                                    {United States}   \n",
-       "ZU5A21A                                      {United States}   \n",
-       "ZZ3249                                       {United States}   \n",
-       "  026L20A                                    {United States}   \n",
-       "ÂK4244                                      {United Kingdom}   \n",
-       "\n",
-       "                                      Countries Intersection  \n",
-       "VAX_LOT                                                       \n",
-       "FE6208     [Austria, Germany, Hungary, Spain, Sweden, Uni...  \n",
-       "039K20A                                      [United States]  \n",
-       "EN6201                       [United Kingdom, United States]  \n",
-       "FD6840     [Austria, Germany, Hungary, Sweden, United Kin...  \n",
-       "FF0680     [Austria, Germany, Hungary, Italy, Netherlands...  \n",
-       "...                                                      ...  \n",
-       "ZTC540693                                    [United States]  \n",
-       "ZU5A21A                                                   []  \n",
-       "ZZ3249                                                    []  \n",
-       "  026L20A                                    [United States]  \n",
-       "ÂK4244                                                    []  \n",
-       "\n",
-       "[5188 rows x 3 columns]"
-      ]
-     },
-     "execution_count": 43,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
-   "source": [
-    "def intersect(series):\n",
-    "    return sorted(set.intersection(*series))\n",
-    "\n",
-    "res['Countries Intersection'] = res.apply(intersect, axis = 'columns')\n",
-    "res"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 46,
-   "id": "a700cb34",
-   "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>Countries Clicked</th>\n",
-       "      <th>Countries Before Deletion</th>\n",
-       "      <th>Countries Intersection</th>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>VAX_LOT</th>\n",
-       "      <th></th>\n",
-       "      <th></th>\n",
-       "      <th></th>\n",
-       "    </tr>\n",
-       "  </thead>\n",
-       "  <tbody>\n",
-       "    <tr>\n",
-       "      <th>FH3253</th>\n",
-       "      <td>{Australia, Malaysia}</td>\n",
-       "      <td>{Austria, Switzerland, Belgium}</td>\n",
-       "      <td>[]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>FE2090</th>\n",
-       "      <td>{United Kingdom, Australia, Germany}</td>\n",
-       "      <td>{Estonia, Spain, France, Denmark, South Africa...</td>\n",
-       "      <td>[]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>FF0884</th>\n",
-       "      <td>{Australia, Malaysia, Japan}</td>\n",
-       "      <td>{Ireland, Denmark}</td>\n",
-       "      <td>[]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>FG2872</th>\n",
-       "      <td>{Malaysia, Brazil}</td>\n",
-       "      <td>{Georgia, United States, Israel}</td>\n",
-       "      <td>[]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>FG3716</th>\n",
-       "      <td>{Australia, Italy, Malaysia}</td>\n",
-       "      <td>{Estonia, Spain, France, Norway, Austria, Denm...</td>\n",
-       "      <td>[]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>...</th>\n",
-       "      <td>...</td>\n",
-       "      <td>...</td>\n",
-       "      <td>...</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>YFL8095</th>\n",
-       "      <td>{Japan}</td>\n",
-       "      <td>{United States}</td>\n",
-       "      <td>[]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>Z07AZ1A</th>\n",
-       "      <td>{Brazil}</td>\n",
-       "      <td>{United States}</td>\n",
-       "      <td>[]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>ZU5A21A</th>\n",
-       "      <td>{}</td>\n",
-       "      <td>{United States}</td>\n",
-       "      <td>[]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>ZZ3249</th>\n",
-       "      <td>{Hungary}</td>\n",
-       "      <td>{United States}</td>\n",
-       "      <td>[]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>ÂK4244</th>\n",
-       "      <td>{Hungary}</td>\n",
-       "      <td>{United Kingdom}</td>\n",
-       "      <td>[]</td>\n",
-       "    </tr>\n",
-       "  </tbody>\n",
-       "</table>\n",
-       "<p>1345 rows × 3 columns</p>\n",
-       "</div>"
-      ],
-      "text/plain": [
-       "                            Countries Clicked  \\\n",
-       "VAX_LOT                                         \n",
-       "FH3253                  {Australia, Malaysia}   \n",
-       "FE2090   {United Kingdom, Australia, Germany}   \n",
-       "FF0884           {Australia, Malaysia, Japan}   \n",
-       "FG2872                     {Malaysia, Brazil}   \n",
-       "FG3716           {Australia, Italy, Malaysia}   \n",
-       "...                                       ...   \n",
-       "YFL8095                               {Japan}   \n",
-       "Z07AZ1A                              {Brazil}   \n",
-       "ZU5A21A                                    {}   \n",
-       "ZZ3249                              {Hungary}   \n",
-       "ÂK4244                              {Hungary}   \n",
-       "\n",
-       "                                 Countries Before Deletion  \\\n",
-       "VAX_LOT                                                      \n",
-       "FH3253                     {Austria, Switzerland, Belgium}   \n",
-       "FE2090   {Estonia, Spain, France, Denmark, South Africa...   \n",
-       "FF0884                                  {Ireland, Denmark}   \n",
-       "FG2872                    {Georgia, United States, Israel}   \n",
-       "FG3716   {Estonia, Spain, France, Norway, Austria, Denm...   \n",
-       "...                                                    ...   \n",
-       "YFL8095                                    {United States}   \n",
-       "Z07AZ1A                                    {United States}   \n",
-       "ZU5A21A                                    {United States}   \n",
-       "ZZ3249                                     {United States}   \n",
-       "ÂK4244                                    {United Kingdom}   \n",
-       "\n",
-       "        Countries Intersection  \n",
-       "VAX_LOT                         \n",
-       "FH3253                      []  \n",
-       "FE2090                      []  \n",
-       "FF0884                      []  \n",
-       "FG2872                      []  \n",
-       "FG3716                      []  \n",
-       "...                        ...  \n",
-       "YFL8095                     []  \n",
-       "Z07AZ1A                     []  \n",
-       "ZU5A21A                     []  \n",
-       "ZZ3249                      []  \n",
-       "ÂK4244                      []  \n",
-       "\n",
-       "[1345 rows x 3 columns]"
-      ]
-     },
-     "execution_count": 46,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
-   "source": [
-    "res[res['Countries Intersection'].apply(lambda s: len(s) == 0)]"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 50,
-   "id": "8fcb32f0",
-   "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>Countries Clicked</th>\n",
-       "      <th>Countries Before Deletion</th>\n",
-       "      <th>Countries Intersection</th>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>VAX_LOT</th>\n",
-       "      <th></th>\n",
-       "      <th></th>\n",
-       "      <th></th>\n",
-       "    </tr>\n",
-       "  </thead>\n",
-       "  <tbody>\n",
-       "    <tr>\n",
-       "      <th>FE6208</th>\n",
-       "      <td>{Germany, Poland, United Kingdom, Australia, B...</td>\n",
-       "      <td>{Lithuania, Spain, Czechia, Greece, Germany, A...</td>\n",
-       "      <td>[Austria, Germany, Hungary, Spain, Sweden, Uni...</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>039K20A</th>\n",
-       "      <td>{Sweden, United States, Spain, France, Austral...</td>\n",
-       "      <td>{United States}</td>\n",
-       "      <td>[United States]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>EN6201</th>\n",
-       "      <td>{Sweden, United States, Spain, France, Austral...</td>\n",
-       "      <td>{United Kingdom, United States}</td>\n",
-       "      <td>[United Kingdom, United States]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>FD6840</th>\n",
-       "      <td>{Sweden, United States, Spain, France, Austral...</td>\n",
-       "      <td>{Lithuania, Cyprus, Czechia, Greece, Germany, ...</td>\n",
-       "      <td>[Austria, Germany, Hungary, Sweden, United Kin...</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>FF0680</th>\n",
-       "      <td>{Sweden, United States, Canada, Brazil, German...</td>\n",
-       "      <td>{Portugal, Sweden, Czechia, France, Finland, G...</td>\n",
-       "      <td>[Austria, Germany, Hungary, Italy, Netherlands...</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>...</th>\n",
-       "      <td>...</td>\n",
-       "      <td>...</td>\n",
-       "      <td>...</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>XXXXXX</th>\n",
-       "      <td>{United States}</td>\n",
-       "      <td>{United States}</td>\n",
-       "      <td>[United States]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>XYZ98765</th>\n",
-       "      <td>{Germany}</td>\n",
-       "      <td>{Germany}</td>\n",
-       "      <td>[Germany]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>Z01AZ1A</th>\n",
-       "      <td>{United States}</td>\n",
-       "      <td>{United States}</td>\n",
-       "      <td>[United States]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>ZTC540693</th>\n",
-       "      <td>{United States}</td>\n",
-       "      <td>{United States}</td>\n",
-       "      <td>[United States]</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>026L20A</th>\n",
-       "      <td>{United States}</td>\n",
-       "      <td>{United States}</td>\n",
-       "      <td>[United States]</td>\n",
-       "    </tr>\n",
-       "  </tbody>\n",
-       "</table>\n",
-       "<p>3843 rows × 3 columns</p>\n",
-       "</div>"
-      ],
-      "text/plain": [
-       "                                           Countries Clicked  \\\n",
-       "VAX_LOT                                                        \n",
-       "FE6208     {Germany, Poland, United Kingdom, Australia, B...   \n",
-       "039K20A    {Sweden, United States, Spain, France, Austral...   \n",
-       "EN6201     {Sweden, United States, Spain, France, Austral...   \n",
-       "FD6840     {Sweden, United States, Spain, France, Austral...   \n",
-       "FF0680     {Sweden, United States, Canada, Brazil, German...   \n",
-       "...                                                      ...   \n",
-       "XXXXXX                                       {United States}   \n",
-       "XYZ98765                                           {Germany}   \n",
-       "Z01AZ1A                                      {United States}   \n",
-       "ZTC540693                                    {United States}   \n",
-       "  026L20A                                    {United States}   \n",
-       "\n",
-       "                                   Countries Before Deletion  \\\n",
-       "VAX_LOT                                                        \n",
-       "FE6208     {Lithuania, Spain, Czechia, Greece, Germany, A...   \n",
-       "039K20A                                      {United States}   \n",
-       "EN6201                       {United Kingdom, United States}   \n",
-       "FD6840     {Lithuania, Cyprus, Czechia, Greece, Germany, ...   \n",
-       "FF0680     {Portugal, Sweden, Czechia, France, Finland, G...   \n",
-       "...                                                      ...   \n",
-       "XXXXXX                                       {United States}   \n",
-       "XYZ98765                                           {Germany}   \n",
-       "Z01AZ1A                                      {United States}   \n",
-       "ZTC540693                                    {United States}   \n",
-       "  026L20A                                    {United States}   \n",
-       "\n",
-       "                                      Countries Intersection  \n",
-       "VAX_LOT                                                       \n",
-       "FE6208     [Austria, Germany, Hungary, Spain, Sweden, Uni...  \n",
-       "039K20A                                      [United States]  \n",
-       "EN6201                       [United Kingdom, United States]  \n",
-       "FD6840     [Austria, Germany, Hungary, Sweden, United Kin...  \n",
-       "FF0680     [Austria, Germany, Hungary, Italy, Netherlands...  \n",
-       "...                                                      ...  \n",
-       "XXXXXX                                       [United States]  \n",
-       "XYZ98765                                           [Germany]  \n",
-       "Z01AZ1A                                      [United States]  \n",
-       "ZTC540693                                    [United States]  \n",
-       "  026L20A                                    [United States]  \n",
-       "\n",
-       "[3843 rows x 3 columns]"
-      ]
-     },
-     "execution_count": 50,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
-   "source": [
-    "res[res['Countries Intersection'].apply(lambda s: len(s) != 0)]"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 44,
-   "id": "7750d699",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "res.to_excel('tmp/res.xlsx')"
+    "countryCountsByBatchcode = getCountryCountsByBatchcodeTable()\n",
+    "countryCountsByBatchcode"
    ]
   }
  ],