Brian Cryer

SQL How To ...

SQL Crib Sheet


These notes are intended to provide a simplified crib sheet (or reminder) on SQL. It is not a tutorial. A number of examples for common types of tasks are provided - but little or no explanation.

SQL - Structured Query Language - is a language understood by most database systems. Except where noted it is believed these SQL statements will work with Microsoft SQL Server, Oracle and MySQL.

Index


Select

Return all records all columns in a table:

select * from TableName

Return all records but only field1 and field2 in a table:

select field1, field2 from TableName

Return field1 for all records in a table with a specific value for field2:

select field1 from TableName where field2=123

Return all records in a table where field1 is one of three possible values:

select * from TableName where field1 in (value1,value2,value3)

Return the number of records in a table:

select count(*) from TheTable

Return the number of records in a table with a specific value for field2:

select count(*) from TableName where field2=123

Simple join:

select * from table1, table2
where table1.field1=table2.fieldA

or

select table1.field1, table2.fieldA from table1, table2
where table1.field2=table2.fieldB

or

select table1.field1, table2.fieldA
from table1 inner join table2 on table1.field2 = table2.fieldB

Select all unique values in field1 from a table (not supported in MSAccess):

select distinct(field1) from TableName

or (works in MSAccess):

select distinct field1 from TableName

To get a count of the number of unique values in a field (not supported in MSAccess):

selet count(distinct field1) from TableName

For MSAccess use:

select count(*) from (select distinct field1 from TableName)

Select all unique values for field1 from a table together with the number of records with that unique value:

select field1, count(*) from TableName
group by field1

Select all unique values for combinations of field1 and field2 from a table together with the number of records with that combination:

select field1, field2, count(*) from TableName
group by field1, field2

Select the number of unique values:

select count(distinct field1) from TableName

Select all duplicate records in a table, where two (or more) records are considered duplicates if they share a common value for a single field:

select field, count(field) from TableName
group by field
having count(*) > 1

Select all duplicate records in a table, where two (or more) records are considered duplicates if they share common values for a pair of fields:

select field1, field2, count(*) from TableName
group by field1, field2
having count(*) > 1

Select similar records, i.e. all records which have duplicate field1 and field2 in a table but with different field3 (i.e. specifying which fields must be the same and which different):

select * from table as A, table as B
where A.field1=B.field1
and A.field2=B.field2
and A.field3<>B.field3;

Note:

  • It is important to specify at least one field which is different between the two records otherwise this query will list a record as being the same as itself.
  • This query will not find duplicate records, i.e. records with every field the same.

Select all records from a table which do not share a common ID with records from a second table:

select * from table1
where field1 not in (select field2 from table2)

Note:

  • Sub-queries are quite slow.
  • Sub-queries are not supported in versions of MySQL prior to MySQL 5, so the above will not work on older versions of MySQL. My thanks to Kevin Bowman for pointing out that MySQL 5 supports sub-queries.

An alternative using a join (which can be much faster):

select table1.* from table1
left join table2 on (table1.field1 = table2.field2)
where table2.field2 is null;

The following method (which has been suggested by Michael Miller) is to use EXISTS. It is much faster on SQL Server than the above (but Michael says it is comparable with the left join technique on Oracle):

select * from table1
where not exists (select field2 from table2 where table2.field2 = table1.field1)

To perform a two way join:

select * from
table1 left join table2 on (table1.field1 = table2.field1),
table1 left join table3 on (table1.field2 = table3.field3)

this has been tested on SQL Server, but not on Oracle or MySql. It does not work with MS-Access.

To combine the results of two queries (be aware that the number and types of fields in both queries must agree):

select * from table1
union select * from table2

To return a value based on the contents of a field. This can be done using either Iif, Decode or Case, depending on the database.

The following works with MSAccess:

select Iif(field1 = 1, 'one', 'not one')
from TableName

This is equivalent to the following on SqlServer:

select Case when field1 = 1 then 'One' else 'Two' End
from TableName

For Oracle use the DECODE function.

To create a new table to hold the results of the select query:

select * into table2 from table1

Be aware that this will fail if table2 exists, and that the new table will be created without any indexes. This isn't supported on MySQL.


Insert

Insert new record into a table:

insert into TableName values (1,2,3)

Insert new record into a table explicitly naming fields:

insert into TableName (field1,field2,field3) values (1,2,3)

Insert new record into a table using values from another table:

insert into TableName (field1,field2,field3)
select fieldA,fieldB,fieldC from SomeTable

MySQL (but not Oracle or SQL Server) allow a single insert statement to insert multiple rows rather than once at a time:

insert into TableName (field1,field2,field3)
values (1,2,3),(4,5,6)


Update

Update all records in a table:

update TableName set field1=2

Update specific records in a table:

update TableName set field1=2 where field1=1

To update more than one field at a time:

update TableName set field1=2, field2=3

Update a field in a table using a value from another table where both records are referenced by a common key - warning, different databases support different syntax!

This works in MS-Access and MySQL (5) but not in SQL Server:

update TableOne
    inner join TableTwo on TableOne.commonID = TableTwo.commonID
    set TableOne.field1 = TableTwo.fieldX

or

This works in MS-Access and MySQL but not in SQL Server:

update TableOne, TableTwo
    set TableOne.field1 = TableTwo.fieldX
    where TableOne.commonID = TableTwo.commonID

or

This works in SQL Server (my thanks to John Lee for this), but not in MS-Access or MySQL:

update tableOne
    set tableOne.field1=tableTwo.fieldX
    from tableOne, tableTwo
    where tableOne.commonID=tableTwo.commonID

Note:

  • MS-Access gives the error "Operation must use an updateable query" if you attempt to use any of the above with a view/query rather than a table. The work around is to copy the data from the query into a temporary table and use the temporary table instead.

Delete

Delete all records in a table (dangerous):

delete from TableName

Delete specific records in a table:

delete from TableName where field1=value

Delete records from one table which do not have a matching field in another table:

delete from TableName where field1 not in
(select field2 from TableTwo)

Keys

Be aware that there are often subtle syntax variations between different database systems. Also other key properties (for example 'clustered') will vary between database systems. Therefore please treat this part of the SQL crib sheet as a guide only.

Create a primary key on a table:

Alter Table TheTable Add Primary Key (field1, field2)

To add an index on a field:

alter table TableName Add Index (field1)

To remove a primary key:

alter table drop primary key


About the author: is a dedicated software developer and webmaster. For his day job he develops websites and desktop applications as well as providing IT services. He moonlights as a technical author and consultant.