home *** CD-ROM | disk | FTP | other *** search
- add the HAVING to GROUP BY in SELECT - DONE 03/30/90
- extend GROUP BY to allow each group to be a range of values
- add ORDER BY to SELECT
- add DELETE command - DONE 04/20/90
- add UPDATE command - DONE 05/04/90
- add subquery to INSERT
- indexes and query optimization
- add null-value logic - four value logic DONE 06/10/91
- implement date-type
- fix so tables produced by evalCols() have proper column types - DONE 02/91
-
- CLOSE DATABASE - close all open tables
-
- develop test suite to test entire functionality
-
- add a shell-type interface extension to command interface
- variables
- aliases
- history
- substitutions
- Get rid of '{' & '}' in evaluating a boolean expression: - DONE 01/28/91
-
- disk and/or RAM based tables, transparent to user - STARTED 04/02/90
- ALL BUT GROUP stuff DONE 04/10/90
- ALL DONE 04/20/90
-
- changed handling of deleted rows, disk access is much faster DONE 09/18/91
-
- Multi-User
-
- Add UNIQUE constraints to CREATE TABLE
- Ex: create table (foo integer, bar integer, unique (foo, bar));
-
- Add some kind of comment delimiters to get_tok()
- DONE in cmd-shell interface 11/07/90
-
- Add command to allow commands to be read in from file and executed
- DONE ??/91
-
- Add SYSTABLES, SYSCOLUMNS, SYSDOMAINS, SYSVIEWS
- ^^^^^^^^^ ^^^^^^^^^^ DONE 7/91
- name name name name
- creator table base type SQL expr
- creator bottom value insertable
- coltype top value updatable
- domain less than deletable
- marks
- length
- scale
- precision
- pkey
- fkey
-
- create above tables
- open/load table - put each column in SYSCOLUMNS, put table in SYSTABLES
- close/drop table - drop rows in SYSCOLUMNS, and SYSTABLES
- crttab.c
- SQLDB:domains - load upon starting, save upon quitting
- should all files be stored in SQLDB? (multi-user definitely)
-
- tablename.cdf column definition file
- tablename.rdf row definition file
-
- BUGS:
- add a check so that a join is not allowed with the same table
- without aliases.
- This is legal: select from foo x, foo y;
- This is not, I think: select from foo, foo;
-
- no keyword test in table definition
-
- no single-value per group test in GROUP BY
-
- GROUP BY doesn't allow use of column numbers (ex. GROUP BY name,2 )
-
- multi-column IN clause:
-
- SELECT * FROM foobar WHERE <col2,col4> IN [
- SELECT col2,col4 FROM footab3 WHERE ...];
-
- or
-
- SELECT * FROM foobar WHERE <col2,col4> IN
- (<pair1a,pair1b>,<pair2a,pair2b>,<pair3a,pair3b>);
-
-
- try this query (gotta implement ORDER BY):
- select sum(1),x.col1,x.col2...
- from mytable x, mytable y
- where x.col1 <= y.col1
- group by x.col1, x.col2...
- order by 1
-
-