Copyright 1998-2008 Curt Hill SQL Queries Yet Another Set of Query Features.

Copyright 1998-2008 Curt Hill

SQL Queries

Yet Another Set of Query Features

Features to Consider• Arithmetic operations• Functions• Set operations • Nested queries


Arithmetic• Numeric fields and constants may

be part of an arithmetic expression – In Where or Select

• These include the standard: * / + -• MySQL and SQL Server accept the

% as modulo• Standard precedence exists and

parentheses may override• There are also other many functions


Arithmetic Example• Considering the age at which a

faculty member started:select f_name, (f_age - f_years) "Starting age" from faculty where f_age - f_years > 40

• The parentheses clarify the statement but are not needed


Functions• There are a number of functions

that may be used anywhere a field name be be used

• The classifications include:– Mathematical functions– String functions


Mathematical• ABS – Absolute value

– Takes one numeric parameter– Usually a field

• Round – See next screen

• Trig functions– Sin, cos, tan are available

• Others– Each DBMS may have others as well


Round• The first parameter is a numeric

expression– This is the only required parameter

• The second is the number of digits to right of decimal– This one is optional

• Thus:Round(s_balance)Round(avg(s_balance),2)

• Oracle has a Trunc, SQL Server Floor


String functions• Substr(field,start,length)

– SQL Server – SubString– Substr(f_name,1,4) gives first four

characters of the name

• Length– Gives the length of a character field– Oracle gives field width for fixed length

fields and actual length for variable length fields

– MySQL disregards trailing blanks for fixed length

– SQL Server: Len


More String Functions• Instr – Find a string in an

expression– Returns position– Instr(crs_title,’Econ’)>0

is true for titles containing this string– Oracle and MySQL

• Lower and Upper– Convert letters to lower or upper case


Word Operators for Where• Between

– Allows use of a range

• Null– Checks for missing value

• Like– A partial rather than full match

• In– Set Membership


Between• Allows a range:Where s Between 25 And 50

• The range is inclusive• This is the same as:Where s >= 25 AND s <= 50

• The values in the range could involve operators or other fields

• This can be negated as well


Example Between• Find the mid-career faculty:select f_name, f_agefrom facultywhere f_age between 30 and 50

• This produces 13 rows


Negated Between• Find those outside a rangeselect f_name, f_age from facultywhere f_age not between 30 and 50

• Orselect f_name, f_age from facultywhere not f_age between 30 and 50

• MySQL must have parentheses around range in last one


Null• Special value that means there is no

value for this field• Not the same as:

– Zero– Blank– Empty string

• Usually obtained by an Insert statement with fewer values than the maximum

• Possible to require a column have a value

• A Primary Key may not be nullCopyright 1998-2008 Curt Hill

Example• Is there a value?select *from studentswhere s_address is Null

• The Not may also be applied before Null or before s_address– MySQL requires:Where not (s_address is Null)in latter case


Like• Expression Like Pattern• Pattern allows two wild cards

– % Any number of any character– _ Any one character– Most other characters appear as

themselves

– SQL Server also allows– [ ] contains allowables– [^ ] any but these


Example• This will find any course with an I

in title:select *from coursewhere crs_title like '% I%'

• Results in 5 rows• Leaving out the trailing % gives

only 3 rows• This should be examined in the

demonstration!Copyright 1998-2008 Curt Hill

Finally on Like• The Not may precede the Like to

negate the whole condition• There is a trailing clause that allows

selection of an escape character:crs_title like ‘%\_%’ escape ‘\’

• Any character may be in the final string

• It is then used to precede one of the wild cards



Set Membership• In is reserved word• Produces a boolean • S in (1,4,5,10)• It is true if s is any of the items• Much easier than many equal

clauses connected with Ors• The set items do not have to be

numeric

Example of In• Find faculty in Math, CIS or MGMT:select *from facultywhere upper(f_dept) in ('MATH', 'CIS', 'MGMT')

• Results in 6 rows• The Not may be applied before the

In or before the Upper– MySQL requires parentheses if before

upper Copyright 1998-2008 Curt Hill


Sub-Queries or Nested Queries

• An in needs a set– Often a constant parenthesized set

• A query produces a set• A nested query is a full query

inside the Where clause of a Select• Generally the sub-query produces

a set of tuples with a single field used in an IN

• Query is inside parentheses


Example Sub-query• MS Divisional classes• Select *From courseWhere crs_dept IN (Select dp_acronym From depart Where dp_division = 'S')


Sub-Query Notes• The sub-query is never seen

– It is resource for the outer query• The sub-query may not use any of

the following clauses:– Into, order by, group by– Why is there need to use these?

• Besides IN there is also– All - All the sub-query meets some test

• The for all ()– Exists – One meets the test– These come from the relational calculus


Another example

• Student who got only As or Bs in CS classesselect s_name, s_idfrom studentswhere 80 <= ALL( Select g_score From grades WHERE s_id = g_naid AND g_dept = 'CS')

• The problem is that this gives way too many: 67

Problem: Nulls• Any student who took no classes in

CS gets an empty table• The comparison becomes 80 <=

Null• The DBMS does not know what to

do with this so calls it true– Which is not what we want

• How do we fix this?


Use an exists• Disallow empty sub-query tablesselect s_name, s_idfrom studentswhere 80 <= ALL( Select g_score From grades WHERE s_id = g_naid AND g_dept = 'CS') AND Exists (Select g_score From grades WHERE s_id = g_naid AND g_dept = 'CS')


What’s wrong with that?• Lots of coding• Will the two sub-queries cost more

than one?• If the query optimizer is good then

no– It will recognize that the two sub-

queries are the same and only do once

• Otherwise yes• There may be other ways as well



Relationship of sub-query to main query

• The scope of names in the parentheses is just the parentheses

• Nothing but the set produced in the sub-query may be used outside

• Things in the outside query may be used inside

• The nested query may be independent

• It may be correlated


Examples• Independent

• It has no values that are related to the main query

• Previous example of finding MS Divisional classes

• Correlated• The main query injects values into the

sub-query• Previous example of finding students

who got an A or B in a CS class

Copyright 1998-2008 Curt Hill SQL Queries Yet Another Set of Query Features.

Documents