Dedupe entries when generating schema tables

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "sinker"
-version = "0.1.1"
+version = "0.1.2"
 description = "Synchronize Postgres to Elasticsearch"
 authors = ["Loren Siebert <[email protected]>"]
 license = "MIT/Apache-2.0"

src/sinker/utils.py

@@ -7,10 +7,14 @@
 
 def generate_schema_tables(view_select_query: str) -> Iterable[str]:
     """
-    Given a view select query, return a list of tables that are referenced in the query.
+    Given a view select query, return a list of unique tables that are referenced in the query
+    in the order they were encountered.
     Skip anything that looks like a function call.
     :param view_select_query: The select query from the view
     """
+    seen: set = set()
     for table_candidate in TABLE_RE.findall(view_select_query):
         if "(" not in table_candidate:
-            yield table_candidate
+            if table_candidate not in seen:
+                seen.add(table_candidate)
+                yield table_candidate

tests/test_generate_schema_tables.py

@@ -5,8 +5,11 @@ def test_generate_schema_tables():
     view_select_query = """select id,
        json_build_object(
                'name', "name",
-               'emailDomains',(select array_agg(split_part(email, '@', 2)) FROM unnest(emails) as email),
+               'otherEmailDomains',(select array_agg(split_part(email, '@', 2)) FROM unnest(emails) as email),
+               'emailDomains', (select array_agg(split_part(value, '@', 2))
+                    from "EmailAddress" EA where "personId"="Person".id),
+               'emailAddresses', (select array_agg(value) from "EmailAddress" EA where "personId"="Person".id),
                ) as "person"
         from "person"
         """
-    assert list(generate_schema_tables(view_select_query)) == ["person"]
+    assert list(generate_schema_tables(view_select_query)) == ["EmailAddress", "person"]