Views, implemented in stepwise fashion up to full functionality. See the section called “Views”. See the section called “CREATE VIEW Syntax”.

Stored procedures currently are implemented, based on the SQL:2003 standard. See Chapter 20, Stored Procedures and Functions.

Elementary cursor support. See the section called “Cursors”.

The ability to specify explicitly for MyISAM tables that an index should be created as an RTREE index. (In MySQL 4.1, RTREE indexes are used internally for geometrical data that use GIS data types, but cannot be created on request.)

Dynamic length rows for MEMORY tables.

Support for Data Dictionary / INFORMATION_SCHEMA.

Add true VARCHAR support (column lengths longer than 255, and no stripping of trailing whitespace). There is already support for this in the MyISAM storage engine, but it is not yet available at the user level.

SHOW COLUMNS FROM tbl_name (used by the mysql client to allow expansions of column names) should not open the table, only the definition file. This will require less memory and be much faster.

Allow DELETE on MyISAM tables to use the record cache. To do this, we need to update the threads record cache when we update the .MYD file.

Better support for MEMORY tables:

Resolving the issue of RENAME TABLE on a table used in an active MERGE table possibly corrupting the table.

FOREIGN KEY support for all table types, not just InnoDB.

Column-level constraints. See the section called “How MySQL Deals with Constraints”.

Online backup with very low performance penalty. The online backup will make it easy to add a new replication slave without taking down the master.

New text based table definition file format (.frm files) and a table cache for table definitions. This will enable us to do faster queries of table structures and do more efficient foreign key support.

Optimize the BIT type to take one bit. (BIT now takes one byte; it is treated as a synonym for TINYINT.)

Add options to the client/server protocol to get progress notes for long running commands.

Implement RENAME DATABASE. To make this safe for all storage engines, it should work as follows:

New internal file interface change. This will make all file handling much more general and make it easier to add extensions like RAID.

Oracle-like CONNECT BY PRIOR to search tree-like (hierarchical) structures.

Add all missing standard SQL and ODBC 3.0 types.

Add SUM(DISTINCT).

INSERT SQL_CONCURRENT and mysqld --concurrent-insert to do a concurrent insert at the end of a table if the table is read-locked.

Allow variables to be updated in UPDATE statements. For example: UPDATE foo SET @a:=a+b,a=@a, b=@a+c.

Change when user variables are updated so that you can use them with GROUP BY, as in the following statement: SELECT id, @a:=COUNT(*), SUM(sum_col)/@a FROM tbl_name GROUP BY id.

Add an IMAGE option to LOAD DATA INFILE to not update TIMESTAMP and AUTO_INCREMENT columns.

Add LOAD DATA INFILE ... UPDATE syntax that works like this:

Make LOAD DATA INFILE understand syntax like this:

LOAD DATA INFILE 'file_name.txt' INTO TABLE tbl_name
     TEXT_FIELDS (text_col1, text_col2, text_col3)
     SET table_col1=CONCAT(text_col1, text_col2),
         table_col3=23
     IGNORE text_col3

This can be used to skip over extra columns in the text file, or update columns based on expressions of the read data.

New functions for working with SET type columns:

If you abort mysql in the middle of a query, you should open another connection and kill the old running query. Alternatively, an attempt should be made to detect this in the server.

Add a storage engine interface for table information so that you can use it as a system table. This would be a bit slow if you requested information about all tables, but very flexible. SHOW INFO FROM tbl_name for basic table information should be implemented.

Allow SELECT a FROM tbl_name1 LEFT JOIN tbl_name2 USING (a); in this case a is assumed to come from tbl_name1.

DELETE and REPLACE options to the UPDATE statement (this will delete rows when a duplicate-key error occurs while updating).

Change the format of DATETIME to store fractions of seconds.

Make it possible to use the new GNU regexp library instead of the current one (the new library should be much faster than the current one).

Add ANY(), EVERY(), and SOME() group functions. In standard SQL, these work only on boolean columns, but we can extend these to work on any columns or expressions by treating a value of zero as FALSE and non-zero values as TRUE.

Fix the type of MAX(column) to be the same as the column type:

mysql> CREATE TABLE t1 (a DATE);
mysql> INSERT INTO t1 VALUES (NOW());
mysql> CREATE TABLE t2 SELECT MAX(a) FROM t1;
mysql> SHOW COLUMNS FROM t2;

Don't allow more than a defined number of threads to run MyISAM recovery at the same time.

Change INSERT INTO ... SELECT to optionally use concurrent inserts.

Add an option to periodically flush key pages for tables with delayed keys if they haven't been used in a while.

Allow join on key parts (optimization issue).

Add a log file analyzer that can extract information about which tables are hit most often, how often multiple-table joins are executed, and so on. This should help users identify areas of table design that could be optimized to execute much more efficient queries.

Return the original column types when doing SELECT MIN(column) ... GROUP BY.

Make it possible to specify long_query_time with a granularity in microseconds.

Link the myisampack code into the server so that it can perform PACK or COMPRESS operations.

Add a temporary key buffer cache during INSERT/DELETE/UPDATE so that we can gracefully recover if the index file gets full.

If you perform an ALTER TABLE on a table that is symlinked to another disk, create temporary tables on that disk.

Implement a DATE/DATETIME type that handles time zone information properly, to make dealing with dates in different time zones easier.

Fix configure so that all libraries (like MyISAM) can be compiled without threads.

Allow user variables as LIMIT arguments; for example, LIMIT @a,@b.

Automatic output from mysql to a Web browser.

LOCK DATABASES (with various options).

Many more variables for SHOW STATUS. Record reads and updates. Selects on a single table and selects with joins. Mean number of tables in selects. Number of ORDER BY and GROUP BY queries.

mysqladmin copy database new-database; this requires a COPY operation to be added to mysqld.

Processlist output should indicate the number of queries/threads.

SHOW HOSTS for printing information about the hostname cache.

Change table names from empty strings to NULL for calculated columns.

Don't use Item_copy_string on numerical values to avoid number-to-string-to-number conversion in case of SELECT COUNT(*)*(id+0) FROM tbl_name GROUP BY id.

Change so that ALTER TABLE doesn't abort clients that execute INSERT DELAYED.

Fix so that when columns are referenced in an UPDATE clause, they contain the old values from before the update started.

Port the MySQL clients to LynxOS.