added country dictionary

benjbaron · Oct 1, 2018 · b9508e2 · b9508e2
1 parent ad6ea71
commit b9508e2
Showing 1 changed file with 4 additions and 1 deletion.
diff --git a/GeoNames Cities Pipeline.ipynb b/GeoNames Cities Pipeline.ipynb
@@ -105,7 +105,7 @@
     "    StructField(\"AC2\", StringType()),\n",
     "    StructField(\"AC3\", StringType()),\n",
     "    StructField(\"AC4\", StringType()),\n",
-    "    StructField(\"POP\", StringType()),\n",
+    "    StructField(\"POP\", IntegerType()),\n",
     "    StructField(\"ALT\", StringType()),\n",
     "    StructField(\"DEM\", StringType()),\n",
     "    StructField(\"TZ\", StringType()),\n",
@@ -307,6 +307,9 @@
     "# User-defined function to transform an array of strings into a string (eq. str.join).\n",
     "array_to_string_udf = udf(lambda x: \",\".join(x))\n",
     "\n",
+    "# Get the countries present in the Postal Codes dataframe\n",
+    "countries = df_pc_filtered.groupby('CC').count().toPandas().set_index('CC').T.to_dict()\n",
+    "\n",
     "print(\"Computing the join...\")\n",
     "for cc, count in countries.items():            \n",
     "    print(\"[.] %s: Processing...\" % cc)\n",