rafa tests exploring weighted_avg

tests/tests_rafa/tests_arrow_vs_duckdb.R

@@ -64,7 +64,8 @@ df_duck_dani <- geocodebr:::geocode2(
   addresses_table = input_df,
   address_fields = fields,
   n_cores = 7,
-  progress = T )
+  progress = T
+  )
 tictoc::toc()
 # 900K: 13 secs
 

tests/tests_rafa/weighted_avg.R

@@ -0,0 +1,32 @@
+I have two tables:
+
+  tableA looks like this:
+
+  id | number
+1 | 100
+2 | 120
+3 | 300
+
+table B looks like this:
+  number | longitude
+100 | -41.1
+100 | -41.101
+150 | -41.3
+300 | -41.2
+
+Write the code in SQL with duckdb to perform left join between tables A and B based on column 'id' so that the logitude of an 'id' will be the average of longitudes in table B weighted by the difference between the numbers in table A and table B
+
+
+
+
+"SELECT
+    A.id,
+    SUM((1/ABS(A.number - B.number) * B.longitude)) / SUM(1/ABS(A.number - B.number)) AS weighted_avg_longitude
+FROM
+    tableA AS A
+LEFT JOIN
+    tableB AS B
+ON
+    A.number = B.number
+GROUP BY
+    A.id;"