Merge pull request #12 from fishtown-analytics/feature/compare-querie…

…s-percent-of-total

Feature/compare queries percent of total

README.md

@@ -15,11 +15,11 @@ relations. It is largely based on the [equality](https://github.com/fishtown-ana
 test in dbt-utils. By default, the generated query returns a summary of audit
 results, like so:
 
-| in_a  | in_b  | count |
-|-------|-------|-------|
-| True  | True  | 6870  |
-| True  | False | 9     |
-| False | True  | 9     |
+| in_a  | in_b  | count | percent_of_total |
+|-------|-------|------:|-----------------:|
+| True  | True  | 6870  | 99.74            |
+| True  | False | 9     | 0.13             |
+| False | True  | 9     | 0.13             |
 
 The generated SQL also contains commented-out SQL that you can use to check
 the rows that do not match perfectly:
@@ -105,7 +105,7 @@ two queries, and summarizes how many records match perfectly (note: a primary
 key is required to match values across the two queries).
 
 | match_status                | count  | percent_of_total |
-|-----------------------------|--------|------------------|
+|-----------------------------|-------:|-----------------:|
 | ✅: perfect match            | 37,721 | 79.03            |
 | ✅: both are null            | 5,789  | 12.13            |
 | 🤷: missing from b          | 25     | 0.05             |

integration_tests/data/expected_results__compare_relations_with_exclude.csv

@@ -1,2 +1,2 @@
-in_a,in_b,count
-True,True,2
+in_a,in_b,count,percent_of_total
+True,True,2,100

integration_tests/data/expected_results__compare_relations_without_exclude.csv

@@ -1,4 +1,4 @@
-in_a,in_b,count
-True,True,1
-True,False,1
-False,True,1
+in_a,in_b,count,percent_of_total
+True,True,1,33.33
+True,False,1,33.33
+False,True,1,33.33

macros/compare_queries.sql

@@ -75,7 +75,11 @@ summary_stats as (
 -- where not (in_a and in_b)
 -- order by {{ primary_key ~ ", " if primary_key is not none }} in_a desc, in_b desc
 
-select * from summary_stats
+select
+    *,
+    round(100.0 * count / sum(count) over (), 2) as percent_of_total
+
+from summary_stats
 order by in_a desc, in_b desc
 
 {% endmacro %}