Add unit argument to haversine distnace macro, and apply conversion for kms (#340)

Author: bastienboutonnet

README.md

@@ -95,7 +95,7 @@ Notes:
 
 ### Date/Time
 #### date_spine ([source](macros/datetime/date_spine.sql))
-This macro returns the sql required to build a date spine. The spine will include the `start_date` (if it is aligned to the `datepart`), but it will not include the `end_date`. 
+This macro returns the sql required to build a date spine. The spine will include the `start_date` (if it is aligned to the `datepart`), but it will not include the `end_date`.
 
 **Usage:**
 
@@ -114,9 +114,12 @@ This macro returns the sql required to build a date spine. The spine will includ
 #### haversine_distance ([source](macros/geo/haversine_distance.sql))
 This macro calculates the [haversine distance](http://daynebatten.com/2015/09/latitude-longitude-distance-sql/) between a pair of x/y coordinates.
 
+Optionally takes a `unit` string parameter ('km' or 'mi') which defaults to miles (imperial system).
+
 **Usage:**
+
 ```
-{{ dbt_utils.haversine_distance(lat1=<float>,lon1=<float>,lat2=<float>,lon2=<float>) }}
+{{ dbt_utils.haversine_distance(lat1=<float>,lon1=<float>,lat2=<float>,lon2=<float>, unit='mi'<string>) }}
 ```
 ---
 ### Schema Tests

integration_tests/data/geo/data_haversine_km.csv

@@ -0,0 +1,2 @@
+lat_1,lon_1,lat_2,lon_2,output
+48.864716,2.349014,52.379189,4.899431,430

integration_tests/data/geo/data_haversine_mi.csv

@@ -0,0 +1,2 @@
+lat_1,lon_1,lat_2,lon_2,output
+48.864716,2.349014,52.379189,4.899431,267

integration_tests/dbt_project.yml

@@ -54,8 +54,8 @@ seeds:
     sql:
       data_events_20180103:
         +schema: events
-    
+
     schema_tests:
       data_test_sequential_timestamps:
         +column_types:
-          my_timestamp: timestamp
+          my_timestamp: timestamp

integration_tests/macros/tests.sql

@@ -1,6 +1,5 @@
 
 {% macro test_assert_equal(model, actual, expected) %}
-
 select count(*) from {{ model }} where {{ actual }} != {{ expected }}
 
 {% endmacro %}

integration_tests/models/geo/schema.yml

@@ -0,0 +1,13 @@
+version: 2
+
+models:
+  - name: test_haversine_distance_km
+    tests:
+      - assert_equal:
+          actual: actual
+          expected: expected
+  - name: test_haversine_distance_mi
+    tests:
+      - assert_equal:
+          actual: actual
+          expected: expected

integration_tests/models/geo/test_haversine_distance_km.sql

@@ -0,0 +1,23 @@
+with data as (
+    select * from {{ ref('data_haversine_km') }}
+), 
+final as (
+    select
+        output as expected,
+        cast(
+            {{
+                dbt_utils.haversine_distance(
+                    lat1='lat_1', 
+                    lon1='lon_1',
+                    lat2='lat_2',
+                    lon2='lon_2',
+                    unit='km'
+                    )
+            }} as numeric
+        ) as actual
+    from data
+)
+select
+    expected,
+    round(actual,0) as actual
+from final

integration_tests/models/geo/test_haversine_distance_mi.sql

@@ -0,0 +1,39 @@
+with data as (
+    select * from {{ ref('data_haversine_mi') }}
+),
+final as (
+    select
+        output as expected,
+        cast(
+            {{
+                dbt_utils.haversine_distance(
+                    lat1='lat_1',
+                    lon1='lon_1',
+                    lat2='lat_2',
+                    lon2='lon_2',
+                    unit='mi'
+                    )
+            }} as numeric
+        ) as actual
+    from data
+
+    union all
+
+    select
+        output as expected,
+        cast(
+            {{
+                dbt_utils.haversine_distance(
+                    lat1='lat_1',
+                    lon1='lon_1',
+                    lat2='lat_2',
+                    lon2='lon_2',
+                    )
+            }} as numeric
+        ) as actual
+    from data
+)
+select
+    expected,
+    round(actual,0) as actual
+from final

macros/geo/haversine_distance.sql

@@ -3,17 +3,49 @@ This calculates the distance between two sets of latitude and longitude.
 The formula is from the following blog post:
 http://daynebatten.com/2015/09/latitude-longitude-distance-sql/
 
-The arguments should be float type. 
+The arguments should be float type.
 #}
 
-{% macro haversine_distance(lat1,lon1,lat2,lon2) -%}
-    {{ return(adapter.dispatch('haversine_distance', packages = dbt_utils._get_utils_namespaces())(lat1,lon1,lat2,lon2)) }}
+{% macro degrees_to_radians(degrees) -%}
+    acos(-1) * {{degrees}} / 180
+{%- endmacro %}
+
+{% macro haversine_distance(lat1, lon1, lat2, lon2, unit='mi') -%}
+    {{ return(adapter.dispatch('haversine_distance', packages = dbt_utils._get_utils_namespaces())(lat1,lon1,lat2,lon2,unit)) }}
 {% endmacro %}
 
-{% macro default__haversine_distance(lat1,lon1,lat2,lon2) -%}
+{% macro default__haversine_distance(lat1, lon1, lat2, lon2, unit='mi') -%}
+{%- if unit == 'mi' %}
+    {% set conversion_rate = 1 %}
+{% elif unit == 'km' %}
+    {% set conversion_rate = 1.60934 %}
+{% else %}
+    {{ exceptions.raise_compiler_error("unit input must be one of 'mi' or 'km'. Got " ~ unit) }}
+{% endif %}
 
-    2 * 3961 * asin(sqrt((sin(radians(({{lat2}} - {{lat1}}) / 2))) ^ 2 +
+    2 * 3961 * asin(sqrt(pow((sin(radians(({{ lat2 }} - {{ lat1 }}) / 2))), 2) +
     cos(radians({{lat1}})) * cos(radians({{lat2}})) *
-    (sin(radians(({{lon2}} - {{lon1}}) / 2))) ^ 2))
+    pow((sin(radians(({{ lon2 }} - {{ lon1 }}) / 2))), 2))) * {{ conversion_rate }}
+
+{%- endmacro %}
+
+
+
+{% macro bigquery__haversine_distance(lat1, lon1, lat2, lon2, unit='mi') -%}
+{% set radians_lat1 = dbt_utils.degrees_to_radians(lat1) %}
+{% set radians_lat2 = dbt_utils.degrees_to_radians(lat2) %}
+{% set radians_lon1 = dbt_utils.degrees_to_radians(lon1) %}
+{% set radians_lon2 = dbt_utils.degrees_to_radians(lon2) %}
+{%- if unit == 'mi' %}
+    {% set conversion_rate = 1 %}
+{% elif unit == 'km' %}
+    {% set conversion_rate = 1.60934 %}
+{% else %}
+    {{ exceptions.raise_compiler_error("unit input must be one of 'mi' or 'km'. Got " ~ unit) }}
+{% endif %}
+    2 * 3961 * asin(sqrt(pow(sin(({{ radians_lat2 }} - {{ radians_lat1 }}) / 2), 2) +
+    cos({{ radians_lat1 }}) * cos({{ radians_lat2 }}) *
+    pow(sin(({{ radians_lon2 }} - {{ radians_lon1 }}) / 2), 2))) * {{ conversion_rate }}
 
 {%- endmacro %}
+