SQL Basics

This page contains the essential commands to manage databases using SQL (Structured Query Language). I use it as a refference for myself. I don't pretend to be an SQL-expert. Others may use this reference freely. If you have any suggestions or comments please e-mail me!
Some commands depend on the type of server used. I experiment with four servers, Microsoft SQL Server, Borland Interbase Server, MySQL Server en Sybase SQL Anywhere. If a command is specific to one server I indicate this with an icon.

1.  indicates that the command works with MS SQL Server.
2.  indicates the command works with Borland InterBase Server.
3. indicates that the command works with MySQL Server.
4. indicates that the command works with Sybase SQL Anywhere.

The commands are written in bold, italicised text (databasenames, fieldnames, values, ...) is up to you to fill in. Note also that I speak of fields and records instaid off rows and coloms. In MS SQL Server manuals I see they use rows and coloms, but I prefer to use fields and records, probably because that was the way in dBase and MS Access, which I used for years.

---

1. List all the databases on the server:

sp_databases 
show databases 

---

2. Select a database to use:

use databasename  

---

3. List tables in the database:

show tables 
sp_help There is another, more complicated, command to retrieve the names of the tables in the database, but where you get a list of ALL tables, including system tables using sp_help, you can specify to view only user defined tables
select * from SYSOBJECTS 
where TYPE = 'U'
order by NAME If you want to see the tables of an InterBase database use the menu "Metadata", "Show..." and select "Tables" in the "View Information on"-dropdown. Click "OK" and you see the tables of the active database. Using ISQL you can use show tables
 To see the tables in Sybase press functionkey F7.

---

4. List fields in a table:

describe tabelname 
sp_help tablename  To see the fields of an InterBase database use the menu "Metadata", "Show..." and select "Tables" in the "View Information on"-dropdown. Type the name of the database in the "Object name"-textfield. Click "OK" and you see the fields of the specified table. Using ISQL you can use show table tablename
 To see the fields in a Sybase table press functionkey F7, select the table and click "Columns".

---

5. Create a database:

create database databasename 
on
(name='databasename_dat',
filename='c:\mssql7\data\databasename_dat.mdf',
size=2mb,
filegrowth=1mb)
log on
(name='databasename_log',
filename='c\mssql7\data\databasename_log.ldf',
size=2mb,
filegrowth=1mb)

---

6. Create a table in the database en store values in it:

create table tablename    
(fieldname1 datatype1, fieldname2 datatype2, ...)
insert into tablename
values(fieldvalue1, fieldvalue2, ...)

---

7. Constraints:

Primary key: constraint constraintname primary key (non)clustered (fieldname1, fieldname2, ...)

Foreign key: constraint constraintname foreign key (fieldname) references tablename(fieldname)

No dulicates: constraint constraintname unique nonclustered (fieldname)

Check: Add values to the fieldnamelist in the 'create table'-command:
(fieldname1 datatype1 check (fieldname1 in ('value1', 'value2', ...)), fieldname2 datatype2, ...)

Default value: Add default value to the fieldlist in the 'create table'-command:
(fieldname1 datatype1, fieldname2 datatype2 default 'value', fieldname3 datatype3, ...)

---

8. Select all records from table:

select * from tablename    
order by fieldnameNote that this command could take some time and put a lot of stress on the processor and memory of the server if you try to view a large database this way. On my server (Pentium MMX 233 Mhz - 64 Mb memory) it took 1:25 minutes to retrieve 25000 records from a database and it gave warnigs of shortage in virtual memory, so please don't try this on a production database (;-). Usualy it is better to select fields and conditions to limit the amount of records and the stress on the server.

---

9. Select set of records from table:

select fieldname1, fieldname2, ... from tablename    
where fieldname = "condition"
order by fieldname
Comparison Operators that can be used in the condition are:

=, !=, <>, <, >, <=, >= and LIKE.

With LIKE you can specify e.g. all the names beginning with S as 'name LIKE "s%"'.
You can also use boolean operators to specify more then one condition (OR, AND, NOT, BETWEEN, IN). With 'BETWEEN' you give a minimum and maximum value to the condition, with 'IN' you can give a list of values.

Example:

select Firstname, Lastname, Department from Employees
where Lastname LIKE "Van%" AND
Badge BETWEEN 121990 and 141990 AND
Department IN ("Sales", "Logistics")
order by Lastname
This statement retrieves all the employees where Lastname begins with "Van", whose badgenumbers are between 121990 and 141990, who work at the Sales- or Logisticsdepartment, and displays their Lastname, Firstname and Dapartment ordered by Lastname.

Note theat the 'ORDER BY' statemant can have more then one fieldname and can also take 'asc' or 'desc' as last argument (for ascending or descnding order).

---

10. Retrieve unique values

select distinct fieldname from table    

---

11. Add records to a table:

insert into tablename (fieldname1, fieldname2, ...)    
values ('value1', 'value2', ...)
If you leave out the fieldnames then the values must match the number of fields in the table. If you only want to add a value to some fields you have to specify them and the values must mach the number of specified fields. The unspecified fields are filled with NULL or the default constraint defined in the table. You could concider to specify defaults like "unknown" or "N/A" to avoid a NULL-value in the field.

---

12. Updating records in a table;

update tablename    
set fieldname1='value1', fieldname2='value2', ...
where fieldname3 = condition
If you ommit the 'where'-statement the update will be performed on ALL the records in the table!

---

13. Deleting a record from a table

delete from tablename    
where fieldname=condition
Be arefull in using this command!! Ommitting the 'where'-statement will erae ALL the records in the table and leave you with an empty table!!!

---

14. Adding a field to the records in a table:

alter table tablename    
add fieldname datatype
The new field is filled with NULL and can be filled using the update-command after adding the field.

---

15. Changing the width of a field

alter table tablename 
alter column fieldname newdatatypealter table tablename   
modify fieldname newdatatype

---

16. Removing field from the records

alter table tablename 
drop column fieldnamealter table tablename   
drop fieldname

---

17. Combine two querries:

select fieldname from tablename
union
select fieldname2 from tablename2
order by fieldname
This union will remove all duplicates. To retain duplicates you have to use 'union all'

---

18. Basic SQL-functions:

select avg(fieldname) from tablename
Returns the arithmeticaverage of the fields.

select count(*) from tablename
where fieldname=condition
Returns the number of records that match the condition.

select max(fieldname) from tablename
Returns the largest value of fieldname.

select min(fieldname) from tablename
Returns the smallest value of fieldname.

select sum(fieldname) from tablename
Returns the summation value of fieldname.

select convert(newdatatype, fieldname) from tablename
converts one datatype into another.

---

19. String functions

ASCII returns ASCII code of leftmost character of a characterexpression.

CHAR converts an ASCII code to a character.

SOUNDEX returns four-digit code used to compare two strings with DIFFERENCE.

DIFFERENCE returns difference between values of two character expressions returned by SOUNDEX.

LEFT returns part of character string, starting at number of character from the left specified in the argument.

LOWER converts uppercase strings to lowercase.

UPPER converts lowercase strings to uppercase.

LTRIM removes leading spaces from a string.

RTRIM removes trailing spaces from a string.

CHARINDEX returns starting position of specified character expression in a string.

PATINDEX rerurns starting position of first occurence of a substring in a string.

REPLICATE returns multiple sets of characters specified in the first argument. The second argument specifies number of sets. (eg. select replicate ('a', 5) returns 'aaaa')

REVERSE returns reverse order of a string of characters.

RIGHT returns part of character string, starting at number of character from the right specified in the argument.

SPACE returns a string of spaces, length specified in argument.

STR converts numeric data to character data.

STUFF inserts a string into another string.

SUBSTRING returns a part of a string (arguments are startpoint and length).

+ (concatenetion) concatenates two or more strings.

---

20. Arithmetic functions:

ACOS, ASIN, ATAN, ATAN2, COS, COT, SIN, TAN, DEGREES, RADIANS, CEILING, FLOOR, EXP, LOG, LOG10, PI(), POWER, ABS, RAND, ROUND, SIGN, SQRT.

---

21. TEXT and IMAGE functions:

SET TEXTSIZE specifies number of bytes displayed for data stored as TEXT or as IMAGE.

TEXTPTR returns pointer to first database page of stoed text.

READTEXT extracts substring from data stored as TEXT or as IMAGE.

TEXTVALID check validity of textpointer.

---

22. Date functions:

DATENAME returns specified part of date as characterstring.

DATEPART returns specified part of date as integer value.

GETDATE returns current date and time.

DATEADD returns value of the date with an additional date interval added.

DATEDIFF returns difference between parts of two specified dates.

---

23. Views:

Note that a view under SQL is the same as a query you create with Access.

23.1 Create views:

create view viewname as
select fieldname1, fieldname2, ... from tablename1, tablename2, ...
where fieldname = conditionYou can't edit a view using SQL. You can use the Enterprise Manager of MS SQL to edit a view or you can delete a view and recreate it with the same name.
You can use alternative names for the columns in a view:
create view viewname (col1, col2, col3, ...)as
select fieldname1, fieldname2, ... from tablename1, tablename2, ...
where fieldname = condition

23.2 Display definition of a view:

sp_helptext viewnameTo prevent the possibility to view the definition of a view you can use encryption:
create view viewname with encryption as
select fieldname1, fieldname2, ... from tablename1, tablename2, ...
where fieldname = condition

23.3 Display associations of a view:

sp_depends viewname

23.4 Delete a view from the database:

drop view viewname1, viewname2, ...

23.5 Insert records through a view:

insert into viewname
values ('value1', 'value2', ...)You can insert rows through the view that DON'T match the WHERE-statement inthe view definition, but then you can't retrieve the new row with the view. If you want to prevent this you can use the check option:
create view viewname as
select fieldname1, fieldname2, ... from tablename1, tablename2, ...
where fieldname = condition with check option

23.6 Delete records through a view:

delete from viewname where fieldname = conditionYou can't delete records through a view that don't match the WHERE-statement in the view definition.

23.7 Updating records through a view:

update viewname set fieldname='value' where fieldname=conditionYou can update a record through a view so that it doesn't match the WHERE-statement anymore. You can't update a view if the updated columns belong to different tables.

---

24. Indexes:

24.1 Simple index:

create index index_name
on table_name (fieldname)

24.2 Unique index:

create unique index index_name
on table_name (fieldname)This kind of index enforces integrity of the table by disallowing duplicete values in the indexed field.

24.3 Clustered index:

create clustered index index_name
on table_name (fieldname)A clustered index forces SQL Server to phisicaly store the table data in the exact order of of the index. This improves the performance of the table. You can only have obne clustered index on a table, the selected fieldname should be choosen carefully and every table should have an clustered index.

24.4 Display index info:

sp_helpindex tablename

24.5 Deleting an index:

drop index table_name.index_name

24.6 Primary and foreign keys:

You can specify a primary key on a table while creating it (see constraint), or you can add a primary key to an existing table by altering it. alter table table_name add constraint constraint_name primary key

24.7 Display primary and foreign keys info:

sp_helpconstraint table_name

24.6 Deleting primary and foreign keys:

alter table table_name drop constraint constraint_name

---

25. Transaction:

A transaction is a series of SQL-statements performed at once, with the ability to undo the changes if something goes wrong during the processing of the statements. A transaction is enclosed in "begin tran" and "commit tran".Example:

begin tran
 update tablename
 set fieldname = newvalue
 where fieldname = condition

 update tablename2
 set fieldname =  newvalue
 where fieldname = condition

 if @@error != 0
 begin
  rollback tran
  print 'Error occured, no rows updated'
  return
 end
commit tran

You can nest one transaction ino another by using named transactions. Be shure however to include a 'commit tran' for every 'begin tran' to avoid leaving a transaction open.
Example:

begin tran tran1_name
 update tablename
 set fieldname = newvalue
 where fieldname = condition

  begin tran tran2_name
   update tablename
   set fieldnname = newvalue
   where fieldname = condition
   if @@error != 0
   begin
    rollback tran tran2_name
    print 'Failed to update tablename'
    return
   end
  commit tran tran2_name
commit tran tran1_name

---

26. Rules:

A rule controls the values that can be entered into a field of a table. To apply a rule to a field you have to take two steps: create the rule and bind the rule to the field.

26.1 Create a rule:

create rule rule_name
as @fieldname in ('value1','value2','value3')

26.2 Bind a rule:

sp_bindrule 'rule_name', 'table_name.field_name'

26.3 Unbind a rule:

sp_unbindrule table_name.field_name

26.4 Drop a rule:

drop rule rule_name
!!You first have to unbind the rule before you can drop it!!To display rule bindings you can use 'sp_help tablename', to display the rules you can use 'sp_helptext rule_name' and to rename a rule you can use 'sp_rename rule_name, new_name'.

---

27. Defaults:

A default is create to provide a value for a field if the user doesn't fill in one. Just like creating rules you have to make to steps to apply ad efault: create it and bind it to a field in a table.

27.1 Create a default:

create default default_name as value

27.2 Bind a default:

sp_bindefault default_name, 'tablename.fieldname'

27.3 Unbind a default:

sp_unbindefault 'tablename.fieldname'

27.4 Drop a default:

drop default default_name
!!You first have to unbind the default before you can drop it!!To display default bindings you can use 'sp_help tablename', to display the default you can use 'sp_helptext default_name' and to rename a default you can use 'sp_rename default_name, new_name'.

---

28. Stored procedures:

Stored procedures are routines or series of SQL-commands that run on the server side. The benefits are performance, because the server typically is the most powerfull computer on the network, and security, because you can control add/change/delete/list operations.

28.1 Create a stored procedure:

create procedure procedure_name
as procedure_commandsTo execute the procedure you use 'exec procedure_name'. You can use parameters in procedures to pass values to the procedure:
create procedure procedure_name (@p1 type, @p2 type, ...)
as insert into table_name
values (@p1, @p2, ...)
Now you can execute this procedure by passing values to it:
exec procedure_name (value1, value2, ...)
You can also use variables in a procedure. You first have to declare them (declare @var_name var_type) and then you can assign a value to it (select @var_name = expression).

28.2 Display a stored procedure:

sp_helptext procedure_name

28.3 Delete a stored procedure:

drop procedure procedure_nameIf you want to alter a procedure, you first have to drop it and then recreate it under the same name.

28.4 Procedure Auto Execution:

You can automaticaly run a procedure every time the server is started by using:sp_makestartup procedure_name
sp_unmakestartup procedure_name removes the procedure from the startupprocedures
sp_helpstartup displays the procedures currently running at startup

28.5 Flow-control statements:

The following statements can be used in stored procedures to control the flow of the procedure:

• if ... else ...
• begin ... end
• while ...
• break
• continue
• print
• goto
• return
• raiserror
• waitfor
• case ... when

---

29. Datatypes:

This is an overview of the different datatypes used by the different servers.

MySQL	Borland InterBase Server	MS SQL Server	Sybase SQL
tinyint	-	tinyint	tinyint
smallint	smallint	smallint	smallint
mediumint	-	-	-
int	int	int	int
bigint	-	-	-
-	numeric	numeric	numeric
decimal	decimal	decimal	decimal
-	-	real	real
float	float	float	float
double	double precision	-	double
date	date	-	date
-	-	smalldatetime	-
datetime	-	datetime	-
timestamp	-	timestamp	timestamp
time	-	-	time
year	-	-	-
char(n)	char(n)	char(n)	char(n)
varchar(n)	varchar(n)	varchar(n)	varchar(n)
-	-	-	long varchar
tinytext	-	-	-
text	-	text	-
-	-	ntext	-
mediumtext	-	-	-
longtext	-	-	-
-	nchar(n)	nchar(n)	-
-	nchar varying(n)	nvarchar(n)	-
-	-	binary(n)	binary(n)
-	-	varbinary(n)	-
-	-	-	long binary
tinyblob	-	-	-
-	-	image	-
blob	blob	-	-
mediumblob	-	-	-
longblob	-	-	-
enum	-	-	-
set	-	-	-
-	-	bit	-
-	-	uniqueidentifier	-
-	-	money	-