Add support for custom sql queries

classes/ezdbischemachecker.php

@@ -57,7 +57,8 @@ public function loadChecksFile( $fileName, $fileFormat )
     public function checkSchema( $returnData=false )
     {
         $violations = array(
-            'FK' => array()
+            'FK' => array(),
+            'Other' => array()
         );
 
         foreach( $this->checks->getForeignKeys() as $def )
@@ -84,6 +85,21 @@ public function checkSchema( $returnData=false )
             }
         }
 
+        foreach( $this->checks->getQueries() as $def )
+        {
+            $violatingRows = $this->countCustomQuery( $def['sql'] );
+
+            if ( $violatingRows )
+            {
+                $def['violatingRowCount'] = $violatingRows;
+                if ( $returnData && $violatingRows )
+                {
+                    $def['violatingRows'] = $this->checkCustomQuery( $def['sql'] );
+                }
+                $violations['Other'][] = $def;
+            }
+        }
+
         return $violations;
     }
 
@@ -167,6 +183,18 @@ public function getFKViolations( $childTable, $childCol, $parentTable, $parentCo
         return $this->db->arrayQuery( $sql );
     }
 
+    public function countCustomQuery( $sql )
+    {
+        $sql = 'SELECT COUNT(*) AS rows FROM ( ' . $sql . ') AS subquery';
+        $results = $this->db->arrayQuery( $sql );
+        return $results[0]['rows'];
+    }
+
+    public function checkCustomQuery( $sql )
+    {
+        return $this->db->arrayQuery( $sql );
+    }
+
     public function getChecks()
     {
         return $this->checks;

classes/ezdbischemachecks.php

@@ -11,12 +11,18 @@
 class ezdbiSchemaChecks
 {
     protected $FK = array();
+    protected $queries = array();
 
     public function getForeignKeys()
     {
         return $this->FK;
     }
 
+    public function getQueries()
+    {
+        return $this->queries;
+    }
+
     public function addForeignKey( $childTable, $childCol, $parentTable, $parentCol, $filter=null, $name='' )
     {
         if ( $name != '' )
@@ -41,6 +47,16 @@ public function addForeignKey( $childTable, $childCol, $parentTable, $parentCol,
         }
     }
 
+    public function addQuery( $sql, $description, $longDesc='' /*, $expectedRows = 0*/ )
+    {
+        $this->queries[] = array(
+            'sql'=> $sql,
+            'description' => $description,
+            'longDesc' => $longDesc,
+            //'expectedRows' => $expectedRows
+        );
+    }
+
     /**
      * Adds checks from another set into this one
      *

classes/schemafileformat/ezdbiiniformat.php

@@ -18,6 +18,7 @@ public function parseFile( $filename )
         $ini = eZINI::instance( $filename );
 
         $checks = new ezdbiSchemaChecks();
+
         foreach( $ini->group( 'ForeignKeys' ) as $table => $value )
         {
             if ( !is_array( $value ) )
@@ -38,6 +39,12 @@ public function parseFile( $filename )
                 }
             }
         }
+
+        foreach( $ini->group( 'CustomQueries' ) as $name => $def )
+        {
+            $checks->addQuery( $def['sql'], str_replace( '_', ' ', $name ), @$def['description'] );
+        }
+
         return $checks;
     }
 
@@ -56,6 +63,17 @@ public function writeFile( $filename, ezdbiSchemaChecks $schemaDef )
             $out .= $def['childTable'] . '[]=' . implode( $this->token, $defs ) . "\n";
         }
 
+        $out .= "\n[CustomQueries]\n";
+        foreach( $schemaDef->getQueries() as $def )
+        {
+            $name = str_replace( ' ', '_', $def['description'] );
+            $out .= $name . '[sql]=' . str_replace( "\n", ' ', $def['sql'] ) . "\n";
+            if ( $def['longDesc'] != '' )
+            {
+                $out .= $name . '[description]=' . str_replace( "\n", ' ', $def['longDesc'] ) . "\n";
+            }
+        }
+
         file_put_contents( $filename, $out );
     }
 }

doc/changelogs/changelog-0.2-to-0.3

@@ -0,0 +1,3 @@
+changelog from version 0.1 to 0.2 - released 2014.3.11
+
+. added data validation for (some) attributes (aka datatypes)

doc/todo

@@ -4,6 +4,8 @@
 
 . start planning for the "bundle" version: load all field defs from services, use built-in validations etc...
 
+. separate checks on data structure from checks on data
+
 
 # attribute checking
 
@@ -20,7 +22,7 @@
 
 . verify if there are more FKs than the ones we know about / some of the current ones to tweak
 
-. add a module/view to display violations in GUI
+. add a module/view to display violations in GUI (custom module in ggsysinfo)
 
 . add per-ez-version inis for FKs and load those instead of generic one
 
@@ -37,3 +39,7 @@
 # more checks
 
 . check data in the db which must fit within ezini values, eg ezflow block types, max content versions, etc
+
+. values we know must be unique but are not
+
+. multi-column foreign keys

extension.xml

@@ -2,7 +2,7 @@
 <software>
     <metadata>
         <name>eZ DB Integrity extension</name>
-        <version>0.1</version>
+        <version>0.3</version>
         <copyright>Copyright (C) 2014 Gaetano Giunta</copyright>
         <license>GNU General Public License v2.0</license>
     </metadata>

ezinfo.php

@@ -6,7 +6,7 @@ static function info()
     {
         return array(
             'Name' => "<a href=\"http://projects.ez.no/ezdbintegrity\">eZ DB Integrity extension</a>",
-            'Version' => "0.1",
+            'Version' => "0.3",
             'Copyright' => "Copyright (C) 2014 Gaetano Giunta",
             'License' => "GNU General Public License v2.0"
         );

settings/ezdbintegrity.ini

@@ -236,3 +236,13 @@ ezworkflow_process[]=memento_key::ezoperation_memento::memento_key
 ezworkflow_process[]=user_id::ezuser::contentobject_id
 ezworkflow_process[]=workflow_id::ezworkflow::id
 ezx_mbpaex[]=contentobject_id::ezcontentobject::id
+
+[CustomQueries]
+urlaliases_without_nodes[sql]=SELECT parent, text_md5, action FROM ezurlalias_ml WHERE ACTION_TYPE = 'eznode' AND is_original = 1 AND substr(action, 8) NOT IN ( SELECT node_id from ezcontentobject_tree )
+nodes_without_urlaliases[sql]=SELECT node_id FROM ezcontentobject_tree WHERE depth > 0 AND node_id NOT IN ( SELECT distinct(substr(action, 8)) as node FROM ezurlalias_ml WHERE ACTION_TYPE = 'eznode' AND is_original = 1 );
+objects_with_many_published_versions[sql]=SELECT contentobject_id, count(*) FROM ezcontentobject_version WHERE status = 1 GROUP BY contentobject_id HAVING COUNT(*) > 1
+ezflow_blocks_with_invalid_zones[sql]=SELECT * FROM ezm_block WHERE zone_id NOT IN ( SELECT DISTINCT ( SUBSTRING( a.data_text, LOCATE( '<zone id="id_', a.data_text ) +13, 32 ) ) AS existing_zone_id FROM ezcontentobject_attribute a, ezcontentobject_version v WHERE a.version = v.version AND a.contentobject_id = v.contentobject_id AND a.data_type_string = 'ezpage' )
+nodes_with_wrong_depth[sql]=SELECT c.node_id FROM ezcontentobject_tree c, ezcontentobject_tree p WHERE c.parent_node_id = p.node_id AND c.depth != ( p.depth + 1 ) AND c.node_id != 1
+objects_with_multiple_main_nodes[sql]=SELECT id, COUNT(DISTINCT main_node_id) FROM o, ezcontentobject_tree n WHERE o.id = n.contentobject_id GROUP BY id, main_node_id HAVING COUNT(DISTINCT main_node_id) > 1
+related_nodes_with_different_objects[sql]=SELECT main_node_id, COUNT(DISTINCT contentobject_id) FROM ezcontentobject_tree GROUP BY main_node_id HAVING COUNT(DISTINCT contentobject_id) > 1
+visible_but_hidden_nodes[sql]=SELECT node_id, is_hidden, is_invisible FROM ezcontentobject_tree WHERE is_hidden=1 AND is_invisible=0