Merge pull request #173 from CartoDB/better-unique-identifiers

Better unique identifiers

CONTRIBUTING.md

@@ -44,7 +44,7 @@ the extension into your test database.
 
 During development the cartodb extension version doesn't change with
 every commit, so testing latest change requires cheating with PostgreSQL
-so to enforce re-load of the scripts. To help with cheating, "make install"
+as to enforce the scripts to reload. To help with cheating, "make install"
 also installs migration scripts to go from "V" to "V"next and from "V"next
 to "V". Example to upgrade a 0.2.0dev version:
 

Makefile

@@ -1,7 +1,7 @@
 # cartodb/Makefile
 
 EXTENSION = cartodb
-EXTVERSION = 0.11.0
+EXTVERSION = 0.11.3
 
 SED = sed
 
@@ -11,6 +11,7 @@ CDBSCRIPTS = \
   scripts-available/CDB_DDLTriggers.sql \
   scripts-available/CDB_ExtensionPost.sql \
   scripts-available/CDB_ExtensionUtils.sql \
+  scripts-available/CDB_Helper.sql \
   $(END)
 
 UPGRADABLE = \
@@ -51,6 +52,9 @@ UPGRADABLE = \
   0.10.1 \
   0.10.2 \
   0.11.0 \
+  0.11.1 \
+  0.11.2 \
+  0.11.3 \
   $(EXTVERSION)dev \
   $(EXTVERSION)next \
   $(END)

NEWS.md

@@ -1,6 +1,23 @@
 next (2015-mm-dd)
 -----------------
 
+0.11.3 (2015-10-27)
+-------------------
+* Added CDB_Helper.sql [#173](https://github.com/CartoDB/cartodb-postgresql/pull/173)
+* Added _CDB_Unique_Identifier for creating UTF8 aware unique identifiers
+* Added _CDB_Unique_Column_Identifier for creating UTF8 aware unique identifiers for columns
+* Added _CDB_Octet_Truncate that truncates text to a certain amount of octets.
+
+0.11.2 (2015-10-19)
+-------------------
+* Fix schema not being specified on pg_get_serial_sequence [#170](https://github.com/CartoDB/cartodb-postgresql/pull/170)
+* Log invalidation function call duration in seconds [#163](https://github.com/CartoDB/cartodb-postgresql/pull/163)
+
+0.11.1 (2015-10-06)
+-------------------
+* Added CDB_DateToNumber(timestamp with time zone) [#169](https://github.com/CartoDB/cartodb-postgresql/pull/169)
+* cartodbfy now discards cartodb_id candidates that contain nulls [#148](https://github.com/CartoDB/cartodb-postgresql/issues/148)
+
 0.11.0 (2015-09-dd)
 -------------------
 * Groups API

scripts-available/CDB_CartodbfyTable.sql

@@ -1,4 +1,5 @@
 -- Depends on:
+--   * CDB_Helper.sql
 --   * CDB_ExtensionUtils.sql
 --   * CDB_TransformToWebmercator.sql
 --   * CDB_TableMetadata.sql
@@ -362,7 +363,7 @@ $$ LANGUAGE PLPGSQL;
 --     As before, this drops all the metadata and geom sync triggers
 --
 -- (2) _CDB_Has_Usable_Primary_ID()
---     Returns TRUE if it can find a unique integer primary key named
+--     Returns TRUE if it can find a unique and not null integer primary key named
 --    'cartodb_id' or can rename an existing key.
 --     Returns FALSE otherwise.
 --
@@ -427,8 +428,9 @@ END;
 $$ LANGUAGE 'plpgsql';
 
 
--- Find a unique relation name in the given schema, starting from the 
--- template given. If the template is already unique, just return it; 
+-- DEPRECATED: Use _CDB_Unique_Identifier since it's UTF8 Safe and length
+-- aware. Find a unique relation name in the given schema, starting from the
+-- template given. If the template is already unique, just return it;
 -- otherwise, append an increasing integer until you find a unique variant.
 CREATE OR REPLACE FUNCTION _CDB_Unique_Relation_Name(schemaname TEXT, relationname TEXT)
 RETURNS TEXT
@@ -439,36 +441,15 @@ DECLARE
   newrelname TEXT;
 BEGIN
 
-  i := 0;
-  newrelname := relationname;
-  LOOP
+  RAISE EXCEPTION '_CDB_Unique_Relation_Name is DEPRECATED. Use _CDB_Unique_Identifier(prefix TEXT, relname TEXT, suffix TEXT, schema TEXT DEFAULT NULL)';
 
-    SELECT c.relname, n.nspname
-    INTO rec
-    FROM pg_class c
-    JOIN pg_namespace n ON c.relnamespace = n.oid
-    WHERE c.relname = newrelname
-    AND n.nspname = schemaname;
-  
-    IF NOT FOUND THEN
-      RETURN newrelname;
-    END IF;
-    
-    i := i + 1;
-    newrelname := relationname || '_' || i;
-  
-    IF i > 100 THEN
-      PERFORM _CDB_Error('looping too far', '_CDB_Unique_Relation_Name');
-    END IF;
-  
-  END LOOP;
-  
 END;
 $$ LANGUAGE 'plpgsql';
 
 
--- Find a unique column name in the given relation, starting from the 
--- column name given. If the column name is already unique, just return it; 
+-- DEPRECATED: Use _CDB_Unique_Column_Identifier since it's UTF8 Safe and length
+-- aware. Find a unique column name in the given relation, starting from the
+-- column name given. If the column name is already unique, just return it;
 -- otherwise, append an increasing integer until you find a unique variant.
 CREATE OR REPLACE FUNCTION _CDB_Unique_Column_Name(reloid REGCLASS, columnname TEXT)
 RETURNS TEXT
@@ -479,32 +460,8 @@ DECLARE
   newcolname TEXT;
 BEGIN
 
-  i := 0;
-  newcolname := columnname;
-  LOOP
+  RAISE EXCEPTION '_CDB_Unique_Column_Name is DEPRECATED. Use _CDB_Unique_Column_Identifier(prefix TEXT, relname TEXT, suffix TEXT, reloid REGCLASS DEFAULT NULL)';
 
-    SELECT a.attname
-    INTO rec
-    FROM pg_class c
-    JOIN pg_attribute a ON a.attrelid = c.oid
-    WHERE NOT a.attisdropped
-    AND a.attnum > 0
-    AND c.oid = reloid
-    AND a.attname = newcolname;
-  
-    IF NOT FOUND THEN
-      RETURN newcolname;
-    END IF;
-    
-    i := i + 1;
-    newcolname := columnname || '_' || i;
-  
-    IF i > 100 THEN
-      PERFORM _CDB_Error('looping too far', '_CDB_Unique_Column_Name');
-    END IF;
-  
-  END LOOP;
-  
 END;
 $$ LANGUAGE 'plpgsql';
 
@@ -551,7 +508,7 @@ BEGIN
         RAISE DEBUG 'CDB(_CDB_Has_Usable_Primary_ID): %', Format('found good ''%s''', const.pkey);
         RETURN true;
 
-      -- Check and see if the column values are unique, 
+      -- Check and see if the column values are unique and not null, 
       -- if they are, we can use this column...
       ELSE
 
@@ -559,13 +516,17 @@ BEGIN
         useable_key := true;
       
         BEGIN
-          sql := Format('ALTER TABLE %s ADD CONSTRAINT %s_unique UNIQUE (%s)', reloid::text, const.pkey, const.pkey);
+          sql := Format('ALTER TABLE %s ADD CONSTRAINT %s_pk PRIMARY KEY (%s)', reloid::text, const.pkey, const.pkey);
           RAISE DEBUG 'CDB(_CDB_Has_Usable_Primary_ID): %', sql;
           EXECUTE sql;
           EXCEPTION      
           -- Failed unique check...
           WHEN unique_violation THEN
-            RAISE NOTICE 'CDB(_CDB_Has_Usable_Primary_ID): %', Format('column %s is not unique', const.pkey);
+            RAISE DEBUG 'CDB(_CDB_Has_Usable_Primary_ID): %', Format('column %s is not unique', const.pkey);
+            useable_key := false;
+          -- Failed not null check...
+          WHEN not_null_violation THEN
+            RAISE DEBUG 'CDB(_CDB_Has_Usable_Primary_ID): %', Format('column %s contains nulls', const.pkey);
             useable_key := false;
           -- Other fatal error
           WHEN others THEN
@@ -574,7 +535,7 @@ BEGIN
   
         -- Clean up test constraint
         IF useable_key THEN
-          PERFORM _CDB_SQL(Format('ALTER TABLE %s DROP CONSTRAINT %s_unique', reloid::text, const.pkey));
+          PERFORM _CDB_SQL(Format('ALTER TABLE %s DROP CONSTRAINT %s_pk', reloid::text, const.pkey));
 
         -- Move non-unique column out of the way
         ELSE
@@ -585,7 +546,7 @@ BEGIN
           PERFORM _CDB_SQL(
             Format('ALTER TABLE %s RENAME COLUMN %s TO %I',
               reloid::text, rec.attname,
-              _CDB_Unique_Column_Name(reloid, const.pkey)),
+              cartodb._CDB_Unique_Column_Identifier(NULL, const.pkey, NULL, reloid)),
             '_CDB_Has_Usable_Primary_ID');
         
         END IF;
@@ -602,7 +563,7 @@ BEGIN
 
       PERFORM _CDB_SQL(
         Format('ALTER TABLE %s RENAME COLUMN %s TO %I',
-                reloid::text, rec.attname, _CDB_Unique_Column_Name(reloid, const.pkey)),
+                reloid::text, rec.attname, cartodb._CDB_Unique_Column_Identifier(NULL, const.pkey, NULL, reloid)),
                 '_CDB_Has_Usable_Primary_ID');
     
     END IF;
@@ -768,7 +729,7 @@ BEGIN
           WHEN others THEN
             IF SQLERRM = 'parse error - invalid geometry' THEN
               text_geom_column := false;
-              str := _CDB_Unique_Column_Name(reloid, r1.attname);
+              str := cartodb._CDB_Unique_Column_Identifier(NULL, r1.attname, NULL, reloid);
               sql := Format('ALTER TABLE %s RENAME COLUMN %s TO %I', reloid::text, r1.attname, str);
               PERFORM _CDB_SQL(sql,'_CDB_Has_Usable_Geom');
               RAISE DEBUG 'CDB(_CDB_Has_Usable_Geom): %', 
@@ -780,7 +741,7 @@ BEGIN
 
       -- Just change its name so we can write a new column into that name.
       ELSE
-        str := _CDB_Unique_Column_Name(reloid, r1.attname);
+        str := cartodb._CDB_Unique_Column_Identifier(NULL, r1.attname, NULL, reloid);
         sql := Format('ALTER TABLE %s RENAME COLUMN %s TO %I', reloid::text, r1.attname, str);
         PERFORM _CDB_SQL(sql,'_CDB_Has_Usable_Geom');
         RAISE DEBUG 'CDB(_CDB_Has_Usable_Geom): %', 
@@ -850,8 +811,7 @@ DECLARE
   destname TEXT;
   destseq TEXT;
   destseqmax INTEGER;
-    
-  salt TEXT := md5(random()::text || now());
+
   copyname TEXT;
 
   column_name_sql TEXT;
@@ -957,19 +917,18 @@ BEGIN
   -- Put the primary key sequence in the right schema
   -- If the new table is not moving, better ensure the sequence name
   -- is unique
-  destseq := relname || '_' || const.pkey || '_seq';
-  destseq := _CDB_Unique_Relation_Name(destschema, destseq);
+  destseq := cartodb._CDB_Unique_Identifier(NULL, relname, '_' || const.pkey || '_seq', destschema);
   destseq := Format('%I.%I', destschema, destseq);
   PERFORM _CDB_SQL(Format('CREATE SEQUENCE %s', destseq), '_CDB_Rewrite_Table');
 
-  -- Salt a temporary table name if we are re-writing in place
+  -- Temporary table name if we are re-writing in place
   -- Note copyname is already escaped and safe to use as identifier
   IF destschema = relschema THEN
-    copyname := Format('%I.%I', destschema, Format('%s_%s', destname, salt));
+    copyname := Format('%I.%I', destschema, cartodb._CDB_Unique_Identifier(NULL, destname, NULL), destschema);
   ELSE
     copyname := Format('%I.%I', destschema, destname);
   END IF;
-  
+
   -- Start building the SQL!
   sql := Format('CREATE TABLE %s AS SELECT ', copyname);
 

scripts-available/CDB_DateToNumber.sql

@@ -13,3 +13,19 @@ RETURN output;
 END;
 $$
 LANGUAGE 'plpgsql' STABLE STRICT;
+
+-- Convert timestamp with time zone to double precision
+--
+CREATE OR REPLACE FUNCTION CDB_DateToNumber(input timestamp with time zone)
+RETURNS double precision AS $$
+DECLARE output double precision;
+BEGIN
+    BEGIN
+        SELECT extract (EPOCH FROM input) INTO output;
+    EXCEPTION WHEN OTHERS THEN
+        RETURN NULL;
+    END;
+RETURN output;
+END;
+$$
+LANGUAGE 'plpgsql' STABLE STRICT;

scripts-available/CDB_Groups.sql

@@ -163,7 +163,7 @@ BEGIN
     group_role := cartodb._CDB_Group_GroupRole(group_name);
     FOR column_name IN EXECUTE 'SELECT COLUMN_NAME FROM INFORMATION_SCHEMA.COLUMNS WHERE TABLE_CATALOG = current_database() AND TABLE_SCHEMA = $1 AND TABLE_NAME = $2 AND COLUMN_DEFAULT LIKE ''nextval%''' USING username, table_name
     LOOP
-        EXECUTE 'SELECT PG_GET_SERIAL_SEQUENCE($1, $2)' USING table_name, column_name INTO sequence_name;
+        EXECUTE format('SELECT PG_GET_SERIAL_SEQUENCE(''%I.%I'', ''%I'')', username, table_name, column_name) INTO sequence_name;
         IF sequence_name IS NOT NULL THEN
           IF do_grant THEN
             -- Here %s is needed since sequence_name has quotes

scripts-available/CDB_Helper.sql

@@ -0,0 +1,156 @@
+-- UTF8 safe and length aware. Find a unique identifier with a given prefix
+-- and/or suffix and withing a schema. If a schema is not specified, the identifier
+-- is guaranteed to be unique for all schemas.
+CREATE OR REPLACE FUNCTION cartodb._CDB_Unique_Identifier(prefix TEXT, relname TEXT, suffix TEXT, schema TEXT DEFAULT NULL)
+RETURNS TEXT
+AS $$
+DECLARE
+  maxlen CONSTANT INTEGER := 63;
+
+  rec RECORD;
+  usedspace INTEGER;
+  ident TEXT;
+  origident TEXT;
+  candrelname TEXT;
+
+  i INTEGER;
+BEGIN
+  -- Accounts for the _XX incremental suffix in case the identifier is taken
+  usedspace := 3;
+  usedspace := usedspace + coalesce(octet_length(prefix), 0);
+  usedspace := usedspace + coalesce(octet_length(suffix), 0);
+
+  candrelname := _CDB_Octet_Truncate(relname, maxlen - usedspace);
+
+  IF candrelname = '' THEN
+    PERFORM _CDB_Error('prefixes are to long to generate a valid identifier', '_CDB_Unique_Identifier');
+  END IF;
+
+  ident := coalesce(prefix, '') || candrelname || coalesce(suffix, '');
+
+  i := 0;
+  origident := ident;
+
+  WHILE i < 100 LOOP
+    IF schema IS NOT NULL THEN
+      SELECT c.relname, n.nspname
+      INTO rec
+      FROM pg_class c
+      JOIN pg_namespace n ON c.relnamespace = n.oid
+      WHERE c.relname = ident
+      AND n.nspname = schema;
+    ELSE
+      SELECT c.relname, n.nspname
+      INTO rec
+      FROM pg_class c
+      JOIN pg_namespace n ON c.relnamespace = n.oid
+      WHERE c.relname = ident;
+    END IF;
+
+    IF NOT FOUND THEN
+      RETURN ident;
+    END IF;
+
+    ident := origident || '_' || i;
+    i := i + 1;
+  END LOOP;
+
+  PERFORM _CDB_Error('looping too far', '_CDB_Unique_Identifier');
+END;
+$$ LANGUAGE 'plpgsql';
+
+
+-- UTF8 safe and length aware. Find a unique identifier for a column with a given prefix
+-- and/or suffix based on colname and within a relation specified via reloid.
+CREATE OR REPLACE FUNCTION cartodb._CDB_Unique_Column_Identifier(prefix TEXT, colname TEXT, suffix TEXT, reloid REGCLASS)
+RETURNS TEXT
+AS $$
+DECLARE
+  maxlen CONSTANT INTEGER := 63;
+
+  rec RECORD;
+  candcolname TEXT;
+  usedspace INTEGER;
+  ident TEXT;
+  origident TEXT;
+
+  i INTEGER;
+BEGIN
+  -- Accounts for the _XX incremental suffix in case the identifier is taken
+  usedspace := 3;
+  usedspace := usedspace + coalesce(octet_length(prefix), 0);
+  usedspace := usedspace + coalesce(octet_length(suffix), 0);
+
+  candcolname := _CDB_Octet_Truncate(colname, maxlen - usedspace);
+
+  IF candcolname = '' THEN
+    PERFORM _CDB_Error('prefixes are to long to generate a valid identifier', '_CDB_Unique_Column_Identifier');
+  END IF;
+
+  ident := coalesce(prefix, '') || candcolname || coalesce(suffix, '');
+
+  i := 0;
+  origident := ident;
+
+  WHILE i < 100 LOOP
+    SELECT a.attname
+    INTO rec
+    FROM pg_class c
+    JOIN pg_attribute a ON a.attrelid = c.oid
+    WHERE NOT a.attisdropped
+    AND a.attnum > 0
+    AND c.oid = reloid
+    AND a.attname = ident;
+
+    IF NOT FOUND THEN
+      RETURN ident;
+    END IF;
+
+    ident := origident || '_' || i;
+    i := i + 1;
+  END LOOP;
+
+  PERFORM _CDB_Error('looping too far', '_CDB_Unique_Column_Identifier');
+END;
+$$ LANGUAGE 'plpgsql';
+
+
+-- Truncates a given string to a max_octets octets taking care
+-- not to leave characters in half. UTF8 safe.
+CREATE OR REPLACE FUNCTION cartodb._CDB_Octet_Truncate(string TEXT, max_octets INTEGER)
+RETURNS TEXT
+AS $$
+DECLARE
+  extcharlen CONSTANT INTEGER := octet_length('ñ');
+
+  expected INTEGER;
+  examined INTEGER;
+  strlen INTEGER;
+
+  i INTEGER;
+BEGIN
+
+  IF max_octets <= 0 THEN
+    RETURN '';
+  ELSIF max_octets >= octet_length(string) THEN
+    RETURN string;
+  END IF;
+
+  strlen := char_length(string);
+
+  expected := char_length(string);
+  examined := octet_length(string);
+
+  IF expected = examined THEN
+    RETURN left(string, max_octets);
+  END IF;
+
+  i := max_octets / extcharlen;
+
+  WHILE octet_length(left(string, i)) <= max_octets LOOP
+    i := i + 1;
+  END LOOP;
+
+  RETURN left(string, (i - 1));
+END;
+$$ LANGUAGE 'plpgsql';

scripts-available/CDB_TableMetadata.sql

@@ -64,6 +64,10 @@ DECLARE
   tabname TEXT;
   rec RECORD;
   found BOOL;
+  function_start timestamptz;
+  function_end timestamptz;
+  function_duration float;
+  log_error_verbosity_value text;
 BEGIN
 
   IF TG_OP = 'UPDATE' or TG_OP = 'INSERT' THEN
@@ -72,6 +76,8 @@ BEGIN
     tabname = OLD.tabname;
   END IF;
 
+  function_start := clock_timestamp();
+ 
   -- Notify table data update
   -- This needs a little bit more of research regarding security issues
   -- see https://github.com/CartoDB/cartodb/pull/241
@@ -105,7 +111,14 @@ BEGIN
     EXIT;
   END LOOP;
   IF NOT found THEN RAISE WARNING 'Missing cdb_invalidate_varnish()'; END IF;
-
+  
+  function_end := clock_timestamp();
+  SELECT extract(epoch from (function_end - function_start)) INTO function_duration;
+  SELECT setting INTO log_error_verbosity_value FROM pg_settings WHERE name='log_error_verbosity';
+  SET log_error_verbosity=TERSE;
+  RAISE LOG 'invalidation_duration: %', function_duration::text;
+  PERFORM 'SET log_error_verbosity= ' || log_error_verbosity_value;
+  
   RETURN NULL;
 END;
 $$

test/CDB_CartodbfyTableTest.sql

@@ -250,7 +250,7 @@ INSERT INTO existing_cartodb_id (cartodb_id, description) VALUES
        (20, 'b'),
        (30, 'c');
 SELECT CDB_CartodbfyTableCheck('existing_cartodb_id', 'Existing cartodb_id values are respected #138');
-SELECT * from existing_cartodb_id;
+SELECT cartodb_id,the_geom,the_geom_webmercator,description,name from existing_cartodb_id;
 DROP TABLE existing_cartodb_id;
 
 -- Table with both the_geom and wkb_geometry
@@ -281,6 +281,44 @@ INSERT INTO many_colliding_columns VALUES (
 SELECT CDB_CartodbfyTableCheck('many_colliding_columns', 'Many colliding columns #141');
 DROP TABLE many_colliding_columns;
 
+-- Table with null cartodb_id
+CREATE TABLE test (
+  cartodb_id integer
+);
+INSERT INTO test VALUES
+  (1), 
+  (2), 
+  (NULL), 
+  (3);
+SELECT CDB_CartodbfyTableCheck('test', 'Table with null cartodb_id #148');
+SELECT cartodb_id, cartodb_id_0 from test;
+DROP TABLE test;
+
+-- Table with non unique cartodb_id
+CREATE TABLE test (
+  cartodb_id integer
+);
+INSERT INTO test VALUES
+  (1), 
+  (2), 
+  (2);
+SELECT CDB_CartodbfyTableCheck('test', 'Table with non unique cartodb_id #148');
+SELECT cartodb_id, cartodb_id_0 from test;
+DROP TABLE test;
+
+-- Table with non unique and null cartodb_id
+CREATE TABLE test (
+  cartodb_id integer
+);
+INSERT INTO test VALUES
+  (1),
+  (2),
+  (NULL),
+  (2);
+SELECT CDB_CartodbfyTableCheck('test', 'Table with non unique and null cartodb_id #148');
+SELECT cartodb_id, cartodb_id_0 from test;
+DROP TABLE test;
+
 
 -- TODO: table with existing custom-triggered the_geom
 

test/CDB_CartodbfyTableTest_expect

@@ -79,5 +79,28 @@ CREATE TABLE
 INSERT 0 1
 Many colliding columns #141 cartodbfied fine
 DROP TABLE
+CREATE TABLE
+INSERT 0 4
+Table with null cartodb_id #148 cartodbfied fine
+1|1
+2|2
+3|
+4|3
+DROP TABLE
+CREATE TABLE
+INSERT 0 3
+Table with non unique cartodb_id #148 cartodbfied fine
+1|1
+2|2
+3|2
+DROP TABLE
+CREATE TABLE
+INSERT 0 4
+Table with non unique and null cartodb_id #148 cartodbfied fine
+1|1
+2|2
+3|
+4|2
+DROP TABLE
 DROP FUNCTION
 DROP FUNCTION

test/CDB_DateToNumberTest.sql

@@ -0,0 +1,2 @@
+SELECT * FROM CDB_DateToNumber('1999-01-08 00:00:00'::timestamp);
+SELECT * FROM CDB_DateToNumber('1999-01-08 00:00:00+05'::timestamp with time zone);

test/CDB_DateToNumberTest_expect

@@ -0,0 +1,2 @@
+915753600
+915735600

test/CDB_HelperTest.sql

@@ -0,0 +1,128 @@
+-- Test unique identifier creation with normal length normal relname
+SELECT * FROM cartodb._CDB_Unique_Identifier(NULL, 'relname', NULL);
+
+-- Test unique identifier creation with prefix with normal length normal relname
+SELECT * FROM cartodb._CDB_Unique_Identifier('prefix_', 'relname', NULL);
+
+-- Test unique identifier creation with suffix with normal length normal relname
+SELECT * FROM cartodb._CDB_Unique_Identifier(NULL, 'relname', '_suffix');
+
+-- Test unique identifier creation with long length normal relname
+SELECT * FROM cartodb._CDB_Unique_Identifier(NULL, 'largolargolargolargolargolargolargolargolargolargolargolargolar', NULL);
+
+-- Test unique identifier creation with prefix with long length normal relname
+SELECT * FROM cartodb._CDB_Unique_Identifier('prefix_', 'largolargolargolargolargolargolargolargolargolargolargolargolar', NULL);
+
+-- Test new identifier is found when name is taken from previous case
+CREATE TABLE prefix_largolargolargolargolargolargolargolargolargolargolar (name text);
+SELECT * FROM cartodb._CDB_Unique_Identifier('prefix_', 'largolargolargolargolargolargolargolargolargolargolargolargolar', NULL);
+DROP TABLE prefix_largolargolargolargolargolargolargolargolargolargolar;
+
+-- Test unique identifier creation with suffix with long length normal relname
+SELECT * FROM cartodb._CDB_Unique_Identifier(NULL, 'largolargolargolargolargolargolargolargolargolargolargolargolar', '_suffix');
+
+-- Test new identifier is found when name is taken from previous case
+CREATE TABLE largolargolargolargolargolargolargolargolargolargolar_suffix (name text);
+SELECT * FROM cartodb._CDB_Unique_Identifier(NULL, 'largolargolargolargolargolargolargolargolargolargolargolargolar', '_suffix');
+DROP TABLE largolargolargolargolargolargolargolargolargolargolar_suffix;
+
+-- Test unique identifier creation with normal length UTF8 relname
+SELECT * FROM cartodb._CDB_Unique_Identifier(NULL, 'piraña', NULL);
+
+-- Test unique identifier creation with prefix with normal length UTF8 relname
+SELECT * FROM cartodb._CDB_Unique_Identifier('prefix_', 'piraña', NULL);
+
+-- Test unique identifier creation with suffix with normal length UTF8 relname
+SELECT * FROM cartodb._CDB_Unique_Identifier(NULL, 'piraña', '_suffix');
+
+-- Test unique identifier creation with long length UTF8 relname
+SELECT * FROM cartodb._CDB_Unique_Identifier(NULL, 'piñaácidpiñaácidpiñaácidpiñaácidpiñaácidpiñaácidpin', NULL);
+
+-- Test unique identifier creation with prefix with long length UTF8 relname
+SELECT * FROM cartodb._CDB_Unique_Identifier('prefix_', 'piñaácidpiñaácidpiñaácidpiñaácidpiñaácidpiñaácidpin', NULL);
+
+-- Test new identifier is found when name is taken from previous case
+CREATE TABLE prefix_piñaácidpiñaácidpiñaácidpiñaácidpiñaácidpi (name text);
+SELECT * FROM cartodb._CDB_Unique_Identifier('prefix_', 'piñaácidpiñaácidpiñaácidpiñaácidpiñaácidpiñaácidpin', NULL);
+DROP TABLE prefix_piñaácidpiñaácidpiñaácidpiñaácidpiñaácidpi;
+
+-- Test unique identifier creation with suffix with long length UTF8 relname
+SELECT * FROM cartodb._CDB_Unique_Identifier(NULL, 'piñaácidpiñaácidpiñaácidpiñaácidpiñaácidpiñaácidpin', '_suffix');
+
+-- Test new identifier is found when name is taken from previous case
+CREATE TABLE piñaácidpiñaácidpiñaácidpiñaácidpiñaácidpi_suffix (name text);
+SELECT * FROM cartodb._CDB_Unique_Identifier(NULL, 'piñaácidpiñaácidpiñaácidpiñaácidpiñaácidpiñaácidpin', '_suffix');
+DROP TABLE piñaácidpiñaácidpiñaácidpiñaácidpiñaácidpi_suffix;
+
+CREATE TABLE test (name text);
+-- Test unique identifier creation with normal length normal colname
+SELECT * FROM cartodb._CDB_Unique_Column_Identifier(NULL, 'colname', NULL, 'test'::regclass);
+
+-- Test unique identifier creation with prefix with normal length normal colname
+SELECT * FROM cartodb._CDB_Unique_Column_Identifier('prefix_', 'colname', NULL, 'test'::regclass);
+
+-- Test unique identifier creation with suffix with normal length normal colname
+SELECT * FROM cartodb._CDB_Unique_Column_Identifier(NULL, 'colname', '_suffix', 'test'::regclass);
+
+-- Test unique identifier creation with long length normal colname
+SELECT * FROM cartodb._CDB_Unique_Column_Identifier(NULL, 'largolargolargolargolargolargolargolargolargolargolargolargolar', NULL, 'test'::regclass);
+
+-- Test unique identifier creation with prefix with long length normal colname
+SELECT * FROM cartodb._CDB_Unique_Column_Identifier('prefix_', 'largolargolargolargolargolargolargolargolargolargolargolargolar', NULL, 'test'::regclass);
+DROP TABLE test;
+
+-- Test new identifier is found when name is taken from previous case
+CREATE TABLE test (prefix_largolargolargolargolargolargolargolargolargolargolar text);
+SELECT * FROM cartodb._CDB_Unique_Column_Identifier('prefix_', 'largolargolargolargolargolargolargolargolargolargolargolargolar', NULL, 'test'::regclass);
+DROP TABLE test;
+
+-- Test unique identifier creation with suffix with long length normal colname
+CREATE TABLE test (name text);
+SELECT * FROM cartodb._CDB_Unique_Column_Identifier(NULL, 'largolargolargolargolargolargolargolargolargolargolargolargolar', '_suffix', 'test'::regclass);
+DROP TABLE test;
+
+-- Test new identifier is found when name is taken from previous case
+CREATE TABLE test (largolargolargolargolargolargolargolargolargolargolar_suffix text);
+SELECT * FROM cartodb._CDB_Unique_Column_Identifier(NULL, 'largolargolargolargolargolargolargolargolargolargolargolargolar', '_suffix', 'test'::regclass);
+DROP TABLE test;
+
+CREATE TABLE test (name text);
+-- Test unique identifier creation with normal length UTF8 colname
+SELECT * FROM cartodb._CDB_Unique_Column_Identifier(NULL, 'piraña', NULL, 'test'::regclass);
+
+-- Test unique identifier creation with prefix with normal length UTF8 colname
+SELECT * FROM cartodb._CDB_Unique_Column_Identifier('prefix_', 'piraña', NULL, 'test'::regclass);
+
+-- Test unique identifier creation with suffix with normal length UTF8 colname
+SELECT * FROM cartodb._CDB_Unique_Column_Identifier(NULL, 'piraña', '_suffix', 'test'::regclass);
+
+-- Test unique identifier creation with long length UTF8 colname
+SELECT * FROM cartodb._CDB_Unique_Column_Identifier(NULL, 'piñaácidpiñaácidpiñaácidpiñaácidpiñaácidpiñaácidpin', NULL, 'test'::regclass);
+
+-- Test unique identifier creation with prefix with long length UTF8 colname
+SELECT * FROM cartodb._CDB_Unique_Column_Identifier('prefix_', 'piñaácidpiñaácidpiñaácidpiñaácidpiñaácidpiñaácidpin', NULL, 'test'::regclass);
+DROP TABLE test;
+
+-- Test new identifier is found when name is taken from previous case
+CREATE TABLE test (prefix_piñaácidpiñaácidpiñaácidpiñaácidpiñaácidpi text);
+SELECT * FROM cartodb._CDB_Unique_Column_Identifier('prefix_', 'piñaácidpiñaácidpiñaácidpiñaácidpiñaácidpiñaácidpin', NULL, 'test'::regclass);
+DROP TABLE test;
+
+-- Test unique identifier creation with suffix with long length UTF8 colname
+CREATE TABLE test (name text);
+SELECT * FROM cartodb._CDB_Unique_Column_Identifier(NULL, 'piñaácidpiñaácidpiñaácidpiñaácidpiñaácidpiñaácidpin', '_suffix', 'test'::regclass);
+DROP TABLE test;
+
+-- Test new identifier is found when name is taken from previous case
+CREATE TABLE test (piñaácidpiñaácidpiñaácidpiñaácidpiñaácidpi_suffix text);
+SELECT * FROM cartodb._CDB_Unique_Column_Identifier(NULL, 'piñaácidpiñaácidpiñaácidpiñaácidpiñaácidpiñaácidpin', '_suffix', 'test'::regclass);
+DROP TABLE test;
+
+-- Test _CDB_Octet_Truncate simple case
+SELECT * FROM cartodb._CDB_Octet_Truncate('piraña', 5);
+
+-- Test _CDB_Octet_Truncate UTF8 case
+SELECT * FROM cartodb._CDB_Octet_Truncate('piraña', 6);
+
+-- Test _CDB_Octet_Truncate UTF8 case
+SELECT * FROM cartodb._CDB_Octet_Truncate('piraña', 7);

test/CDB_HelperTest_expect

@@ -0,0 +1,59 @@
+relname
+prefix_relname
+relname_suffix
+largolargolargolargolargolargolargolargolargolargolargolargo
+prefix_largolargolargolargolargolargolargolargolargolargolar
+CREATE TABLE
+prefix_largolargolargolargolargolargolargolargolargolargolar_0
+DROP TABLE
+largolargolargolargolargolargolargolargolargolargolar_suffix
+CREATE TABLE
+largolargolargolargolargolargolargolargolargolargolar_suffix_0
+DROP TABLE
+piraña
+prefix_piraña
+piraña_suffix
+piñaácidpiñaácidpiñaácidpiñaácidpiñaácidpiñaácid
+prefix_piñaácidpiñaácidpiñaácidpiñaácidpiñaácidpi
+CREATE TABLE
+prefix_piñaácidpiñaácidpiñaácidpiñaácidpiñaácidpi_0
+DROP TABLE
+piñaácidpiñaácidpiñaácidpiñaácidpiñaácidpi_suffix
+CREATE TABLE
+piñaácidpiñaácidpiñaácidpiñaácidpiñaácidpi_suffix_0
+DROP TABLE
+CREATE TABLE
+colname
+prefix_colname
+colname_suffix
+largolargolargolargolargolargolargolargolargolargolargolargo
+prefix_largolargolargolargolargolargolargolargolargolargolar
+DROP TABLE
+CREATE TABLE
+prefix_largolargolargolargolargolargolargolargolargolargolar_0
+DROP TABLE
+CREATE TABLE
+largolargolargolargolargolargolargolargolargolargolar_suffix
+DROP TABLE
+CREATE TABLE
+largolargolargolargolargolargolargolargolargolargolar_suffix_0
+DROP TABLE
+CREATE TABLE
+piraña
+prefix_piraña
+piraña_suffix
+piñaácidpiñaácidpiñaácidpiñaácidpiñaácidpiñaácid
+prefix_piñaácidpiñaácidpiñaácidpiñaácidpiñaácidpi
+DROP TABLE
+CREATE TABLE
+prefix_piñaácidpiñaácidpiñaácidpiñaácidpiñaácidpi_0
+DROP TABLE
+CREATE TABLE
+piñaácidpiñaácidpiñaácidpiñaácidpiñaácidpi_suffix
+DROP TABLE
+CREATE TABLE
+piñaácidpiñaácidpiñaácidpiñaácidpiñaácidpi_suffix_0
+DROP TABLE
+pira
+pirañ
+piraña