From f4e84d93badcd30f75c42515f86ad91e8e27a192 Mon Sep 17 00:00:00 2001
From: Artem Zuikov <chertus@gmail.com>
Date: Thu, 20 Aug 2020 21:51:22 +0300
Subject: [PATCH] fix indirect distinct with less columns

---
 src/Interpreters/TreeOptimizer.cpp                   |  8 ++++++++
 .../01455_duplicate_distinct_optimization.reference  | 12 ++++++++++++
 .../01455_duplicate_distinct_optimization.sql        |  1 +
 3 files changed, 21 insertions(+)

diff --git a/src/Interpreters/TreeOptimizer.cpp b/src/Interpreters/TreeOptimizer.cpp
index 3bded23708f..72f4e961c78 100644
--- a/src/Interpreters/TreeOptimizer.cpp
+++ b/src/Interpreters/TreeOptimizer.cpp
@@ -356,6 +356,7 @@ std::unordered_set<String> getDistinctNames(const ASTSelectQuery & select)
 
     std::unordered_set<String> names;
     std::unordered_set<String> implicit_distinct;
+    std::unordered_set<String> used_implicits;
 
     if (!select.distinct)
     {
@@ -382,6 +383,9 @@ std::unordered_set<String> getDistinctNames(const ASTSelectQuery & select)
                     names.insert(name);
                 else
                     names.insert(alias);
+
+                if (!select.distinct)
+                    used_implicits.insert(name);
             }
         }
         else if (select.distinct && !alias.empty())
@@ -391,6 +395,10 @@ std::unordered_set<String> getDistinctNames(const ASTSelectQuery & select)
         }
     }
 
+    /// SELECT a FROM (SELECT DISTINCT a, b FROM ...)
+    if (!select.distinct && used_implicits.size() != implicit_distinct.size())
+        return {};
+
     return names;
 }
 
diff --git a/tests/queries/0_stateless/01455_duplicate_distinct_optimization.reference b/tests/queries/0_stateless/01455_duplicate_distinct_optimization.reference
index 74addf2fcfa..c5a06ca0cd3 100644
--- a/tests/queries/0_stateless/01455_duplicate_distinct_optimization.reference
+++ b/tests/queries/0_stateless/01455_duplicate_distinct_optimization.reference
@@ -97,6 +97,18 @@ FROM
         FROM numbers(100)
     )
 )
+SELECT DISTINCT a
+FROM 
+(
+    SELECT a
+    FROM 
+    (
+        SELECT DISTINCT 
+            number % 2 AS a,
+            number % 3 AS b
+        FROM numbers(100)
+    )
+)
 SELECT DISTINCT number
 FROM 
 (
diff --git a/tests/queries/0_stateless/01455_duplicate_distinct_optimization.sql b/tests/queries/0_stateless/01455_duplicate_distinct_optimization.sql
index ded170fdf95..10937819de2 100644
--- a/tests/queries/0_stateless/01455_duplicate_distinct_optimization.sql
+++ b/tests/queries/0_stateless/01455_duplicate_distinct_optimization.sql
@@ -11,6 +11,7 @@ ANALYZE SELECT DISTINCT a FROM (SELECT DISTINCT a FROM (SELECT DISTINCT number %
 ANALYZE SELECT DISTINCT a FROM (SELECT DISTINCT a, b FROM (SELECT DISTINCT number % 2 AS a, number % 3 AS b FROM numbers(100)));
 ANALYZE SELECT DISTINCT a, b FROM (SELECT DISTINCT b, a FROM (SELECT DISTINCT number a, number b FROM numbers(1)));
 ANALYZE SELECT DISTINCT a, b FROM (SELECT b, a, a + b FROM (SELECT DISTINCT number % 2 AS a, number % 3 AS b FROM numbers(100)));
+ANALYZE SELECT DISTINCT a FROM (SELECT a FROM (SELECT DISTINCT number % 2 AS a, number % 3 AS b FROM numbers(100)));
 ANALYZE SELECT DISTINCT number FROM (SELECT DISTINCT number FROM numbers(1)) t1 CROSS JOIN numbers(2) t2;
 ANALYZE SELECT DISTINCT number FROM (SELECT DISTINCT number FROM numbers(1) t1 CROSS JOIN numbers(2) t2);