add cnf rewriting to system schema queries

Signed-off-by: Harshit Gangal <harshit@planetscale.com>
vitessio · Jul 19, 2021 · 38f506b · 38f506b
1 parent 260e67c
commit 38f506b
Show file tree

Hide file tree

Showing 4 changed files with 49 additions and 34 deletions.
diff --git a/go/vt/vtgate/engine/route.go b/go/vt/vtgate/engine/route.go
@@ -21,6 +21,7 @@ import (
 	"fmt"
 	"sort"
 	"strconv"
+	"strings"
 	"time"
 
 	"vitess.io/vitess/go/vt/log"
@@ -800,17 +801,12 @@ func (route *Route) description() PrimitiveDescription {
 		other["SysTableTableSchema"] = sysTabSchema
 	}
 	if len(route.SysTableTableName) != 0 {
-		idx := 0
-		sysTableName := "["
+		var sysTableName []string
 		for k, v := range route.SysTableTableName {
-			if idx != 0 {
-				sysTableName += ", "
-			}
-			sysTableName += k + ":" + v.String()
-			idx++
+			sysTableName = append(sysTableName, k+":"+v.String())
 		}
-		sysTableName += "]"
-		other["SysTableTableName"] = sysTableName
+		sort.Strings(sysTableName)
+		other["SysTableTableName"] = "[" + strings.Join(sysTableName, ", ") + "]"
 	}
 	orderBy := GenericJoin(route.OrderBy, orderByToString)
 	if orderBy != "" {

diff --git a/go/vt/vtgate/planbuilder/route_planning.go b/go/vt/vtgate/planbuilder/route_planning.go
@@ -41,11 +41,28 @@ func gen4Planner(_ string) func(sqlparser.Statement, *sqlparser.ReservedVars, Co
 		if !ok {
 			return nil, vterrors.Errorf(vtrpcpb.Code_UNIMPLEMENTED, "%T not yet supported", stmt)
 		}
-		return newBuildSelectPlan(sel, reservedVars, vschema)
+
+		getPlan := func(sel *sqlparser.Select) (logicalPlan, error) {
+			return newBuildSelectPlan(sel, reservedVars, vschema)
+		}
+
+		plan, err := getPlan(sel)
+		if err != nil {
+			return nil, err
+		}
+
+		if shouldRetryWithCNFRewriting(plan) {
+			// by transforming the predicates to CNF, the planner will sometimes find better plans
+			primitive := rewriteToCNFAndReplan(stmt, getPlan)
+			if primitive != nil {
+				return primitive, nil
+			}
+		}
+		return plan.Primitive(), nil
 	}
 }
 
-func newBuildSelectPlan(sel *sqlparser.Select, reservedVars *sqlparser.ReservedVars, vschema ContextVSchema) (engine.Primitive, error) {
+func newBuildSelectPlan(sel *sqlparser.Select, reservedVars *sqlparser.ReservedVars, vschema ContextVSchema) (logicalPlan, error) {
 
 	directives := sqlparser.ExtractCommentDirectives(sel.Comments)
 	if len(directives) > 0 {
@@ -100,7 +117,8 @@ func newBuildSelectPlan(sel *sqlparser.Select, reservedVars *sqlparser.ReservedV
 	if err := plan.WireupGen4(semTable); err != nil {
 		return nil, err
 	}
-	return plan.Primitive(), nil
+
+	return plan, nil
 }
 
 func optimizeQuery(opTree abstract.Operator, reservedVars *sqlparser.ReservedVars, semTable *semantics.SemTable, vschema ContextVSchema) (joinTree, error) {

diff --git a/go/vt/vtgate/planbuilder/testdata/filter_cases.txt b/go/vt/vtgate/planbuilder/testdata/filter_cases.txt
@@ -1959,25 +1959,6 @@ Gen4 plan same as above
   }
 }
 
-# able to isolate table_schema value even when hidden inside of ORs
-"SELECT * FROM INFORMATION_SCHEMA.TABLES WHERE (TABLE_SCHEMA = 'ks' and other_column = 42) OR (TABLE_SCHEMA = 'ks' and foobar = 'value')"
-{
-  "QueryType": "SELECT",
-  "Original": "SELECT * FROM INFORMATION_SCHEMA.TABLES WHERE (TABLE_SCHEMA = 'ks' and other_column = 42) OR (TABLE_SCHEMA = 'ks' and foobar = 'value')",
-  "Instructions": {
-    "OperatorType": "Route",
-    "Variant": "SelectDBA",
-    "Keyspace": {
-      "Name": "main",
-      "Sharded": false
-    },
-    "FieldQuery": "select * from INFORMATION_SCHEMA.`TABLES` where 1 != 1",
-    "Query": "select * from INFORMATION_SCHEMA.`TABLES` where TABLE_SCHEMA = :__vtschemaname and (other_column = 42 or TABLE_SCHEMA = 'ks') and (other_column = 42 or foobar = 'value')",
-    "SysTableTableSchema": "[VARBINARY(\"ks\")]",
-    "Table": "INFORMATION_SCHEMA.`TABLES`"
-  }
-}
-
 # solving LIKE query with a CFC prefix vindex
 "select c2 from cfc_vindex_col where c1 like 'A%'"
 {

diff --git a/go/vt/vtgate/planbuilder/testdata/systemtables_cases.txt b/go/vt/vtgate/planbuilder/testdata/systemtables_cases.txt
@@ -234,7 +234,7 @@ Gen4 plan same as above
     },
     "FieldQuery": "select KCU.DELETE_RULE, S.UPDATE_RULE from INFORMATION_SCHEMA.KEY_COLUMN_USAGE as KCU, INFORMATION_SCHEMA.REFERENTIAL_CONSTRAINTS as RC, INFORMATION_SCHEMA.K as S where 1 != 1",
     "Query": "select KCU.DELETE_RULE, S.UPDATE_RULE from INFORMATION_SCHEMA.KEY_COLUMN_USAGE as KCU, INFORMATION_SCHEMA.REFERENTIAL_CONSTRAINTS as RC, INFORMATION_SCHEMA.K as S where S.TABLE_SCHEMA = :__vtschemaname and S.TABLE_NAME = :S_TABLE_NAME and KCU.TABLE_SCHEMA = :__vtschemaname and KCU.TABLE_NAME = :KCU_TABLE_NAME and KCU.CONSTRAINT_NAME = RC.CONSTRAINT_NAME order by KCU.CONSTRAINT_NAME asc, KCU.COLUMN_NAME asc",
-    "SysTableTableName": "[S_TABLE_NAME:VARBINARY(\"sc\"), KCU_TABLE_NAME:VARBINARY(\"data_type_table\")]",
+    "SysTableTableName": "[KCU_TABLE_NAME:VARBINARY(\"data_type_table\"), S_TABLE_NAME:VARBINARY(\"sc\")]",
     "SysTableTableSchema": "[VARBINARY(\"test\"), VARBINARY(\"test\")]",
     "Table": "INFORMATION_SCHEMA.K, INFORMATION_SCHEMA.KEY_COLUMN_USAGE, INFORMATION_SCHEMA.REFERENTIAL_CONSTRAINTS"
   }
@@ -345,7 +345,7 @@ Gen4 plan same as above
     },
     "FieldQuery": "select fk.referenced_table_name as to_table, fk.referenced_column_name as primary_key, fk.column_name as `column`, fk.constraint_name as `name`, rc.update_rule as on_update, rc.delete_rule as on_delete from information_schema.referential_constraints as rc, information_schema.key_column_usage as fk where 1 != 1",
     "Query": "select fk.referenced_table_name as to_table, fk.referenced_column_name as primary_key, fk.column_name as `column`, fk.constraint_name as `name`, rc.update_rule as on_update, rc.delete_rule as on_delete from information_schema.referential_constraints as rc, information_schema.key_column_usage as fk where rc.constraint_schema = database() and rc.table_name = :rc_table_name and fk.referenced_column_name is not null and fk.table_schema = database() and fk.table_name = :fk_table_name and rc.constraint_schema = fk.constraint_schema and rc.constraint_name = fk.constraint_name",
-    "SysTableTableName": "[rc_table_name:VARBINARY(\":vtg1\"), fk_table_name:VARBINARY(\":vtg1\")]",
+    "SysTableTableName": "[fk_table_name:VARBINARY(\":vtg1\"), rc_table_name:VARBINARY(\":vtg1\")]",
     "Table": "information_schema.key_column_usage, information_schema.referential_constraints"
   }
 }
@@ -422,7 +422,7 @@ Gen4 plan same as above
     },
     "FieldQuery": "select fk.referenced_table_name as to_table, fk.referenced_column_name as primary_key, fk.column_name as `column`, fk.constraint_name as `name`, rc.update_rule as on_update, rc.delete_rule as on_delete from information_schema.referential_constraints as rc, information_schema.key_column_usage as fk where 1 != 1",
     "Query": "select fk.referenced_table_name as to_table, fk.referenced_column_name as primary_key, fk.column_name as `column`, fk.constraint_name as `name`, rc.update_rule as on_update, rc.delete_rule as on_delete from information_schema.referential_constraints as rc, information_schema.key_column_usage as fk where rc.constraint_schema = :__vtschemaname and rc.table_name = :rc_table_name and fk.referenced_column_name is not null and fk.table_schema = :__vtschemaname and fk.table_name = :fk_table_name and rc.constraint_schema = fk.constraint_schema and rc.constraint_name = fk.constraint_name",
-    "SysTableTableName": "[rc_table_name:VARBINARY(\"table_name\"), fk_table_name:VARBINARY(\"table_name\")]",
+    "SysTableTableName": "[fk_table_name:VARBINARY(\"table_name\"), rc_table_name:VARBINARY(\"table_name\")]",
     "SysTableTableSchema": "[VARBINARY(\"table_schema\"), VARBINARY(\"table_schema\")]",
     "Table": "information_schema.key_column_usage, information_schema.referential_constraints"
   }
@@ -738,3 +738,23 @@ Gen4 plan same as above
   }
 }
 Gen4 plan same as above
+
+# able to isolate table_schema value even when hidden inside of ORs
+"SELECT * FROM INFORMATION_SCHEMA.TABLES WHERE (TABLE_SCHEMA = 'ks' and other_column = 42) OR (TABLE_SCHEMA = 'ks' and foobar = 'value')"
+{
+  "QueryType": "SELECT",
+  "Original": "SELECT * FROM INFORMATION_SCHEMA.TABLES WHERE (TABLE_SCHEMA = 'ks' and other_column = 42) OR (TABLE_SCHEMA = 'ks' and foobar = 'value')",
+  "Instructions": {
+    "OperatorType": "Route",
+    "Variant": "SelectDBA",
+    "Keyspace": {
+      "Name": "main",
+      "Sharded": false
+    },
+    "FieldQuery": "select * from INFORMATION_SCHEMA.`TABLES` where 1 != 1",
+    "Query": "select * from INFORMATION_SCHEMA.`TABLES` where TABLE_SCHEMA = :__vtschemaname and (other_column = 42 or TABLE_SCHEMA = 'ks') and (other_column = 42 or foobar = 'value')",
+    "SysTableTableSchema": "[VARBINARY(\"ks\")]",
+    "Table": "INFORMATION_SCHEMA.`TABLES`"
+  }
+}
+Gen4 plan same as above