1. What is SQL?

1. Structured Query Language (SQL) is a specialized language for

updating, deleting, and requesting information from databases.
2. SQL is responsible for querying and editing information stored in
a certain database management system.
3. !e original "ersion called S#Q$#L (structured Englis! query
language) %as designed by an &'( researc! center in 1)*+ and
1)*,.
2. Types of SQL statements
!ere are four basic types of SQL statements-
.ata definition language (..L) statements
.ata manipulation language (.(L) statements
.ynamic SQL statements
(iscellaneous statements
3. What are the DDL, DML an D!L "ommans?
DDL
Data Definition Lan#ua#e (..L) statements are used to define t!e
database structure or sc!ema. Some e/amples-
o 01#2# 3 to create ob4ects in t!e database
o 2L#1 3 alters t!e structure of t!e database
o .156 3 delete ob4ects from t!e database
o 1$702# 3 remo"e all records from a table, including all spaces
allocated for t!e records are remo"ed
o 05((#7 3 add comments to t!e data dictionary
o 1#72(# 3 rename an ob4ect
DML
Data Manipulation Lan#ua#e (.(L) statements are used for
managing data %it!in sc!ema ob4ects. Some e/amples-
o S#L#0 3 retrie"e data from t!e a database
o &7S#1 3 insert data into a table
o $6.2# 3 updates e/isting data %it!in a table
o .#L## 3 deletes all records from a table, t!e space for t!e
records remain
o (#18# 3 $6S#1 operation (insert or update)
o 02LL 3 call a 6L9SQL or :a"a subprogram
o #;6L2&7 6L27 3 e/plain access pat! to data
o L50< 2'L# 3 control concurrency
D!L
Data !ontrol Lan#ua#e $D!L% statements. Some
e&amples'
o 8127 3 gi"es user=s access pri"ileges to database
o 1#>5<# 3 %it!dra% access pri"ileges gi"en %it! t!e 8127
command
T!L
Transa"tion !ontrol (0L) statements are used to manage t!e
c!anges made by .(L statements. &t allo%s statements to be grouped
toget!er into logical transactions.
o 05((& 3 sa"e %or? done
o S2>#65&7 3 identify a point in a transaction to %!ic! you can
later roll bac?
o 15LL'20< 3 restore database to original since t!e last 05((&
o S# 127S20&57 3 0!ange transaction options li?e isolation
le"el and %!at rollbac? segment to use
(. What is )ormali*ation?
7ormalization is t!e process of efficiently organizing data in a
database. !ere are t%o goals of t!e normalization process-
eliminating redundant data (for e/ample, storing t!e same data in
more t!an one table) and ensuring data dependencies ma?e sense
(only storing related data in a table). 'ot! of t!ese are %ort!y goals
as t!ey reduce t!e amount of space a database consumes and ensure
t!at data is logically stored.
+irst )ormal +orm $1)+%
@irst normal form (17@) sets t!e "ery basic rules for an organized
database-
#liminate duplicati"e columns from t!e same table.
0reate separate tables for eac! group of related data and
identify eac! ro% %it! a unique column or set of columns
(t!e primary ?ey).
Se"on )ormal +orm $2)+%
Second normal form (27@) furt!er addresses t!e concept of remo"ing
duplicati"e data-
(eet all t!e requirements of t!e first normal form.
1emo"e subsets of data t!at apply to multiple ro%s of a table
and place t!em in separate tables.
0reate relations!ips bet%een t!ese ne% tables and t!eir
predecessors t!roug! t!e use of foreign ?eys.
Thir )ormal +orm $3)+%
!ird normal form (37@) goes one large step furt!er-
(eet all t!e requirements of t!e second normal form.
1emo"e columns t!at are not dependent upon t!e primary ?ey.
+ourth )ormal +orm $()+%
@inally, fourt! normal form (+7@) !as one additional requirement-
(eet all t!e requirements of t!e t!ird normal form.
2 relation is in +7@ if it !as no multi3"alued dependencies.
,. A!at is constraintsB
SQL Ser"er userCs constraints to enforce limitations on t!e data t!at
can be entered into a particular column in table. !ere are follo%ing
types of constraints.
$nique, .efault, 0!ec?, 6rimary <ey, @oreign <ey, 7ot 7ull.
Synta/-
Synta/
2L#1 2'L# table37ame
D
2.. 05L$(7 column3definition E
2.. 057S12&7 clause E
.156 D 61&(21F <#F E @51#&87 <#F constraint3name E $7&Q$#
constraint3name E 0G#0< constraint3name E 057S12&7
constraint3name H
2L#1 column3alteration E
L50<S&I# D 15A E 2'L# H
H
,. Types of "onstraints in sql
(icrosoft SQL Ser"er supports t!e follo%ing constraints-
61&(21F <#F
$7&Q$#
@51#&87 <#F
0G#0<
75 7$LL
2 61&(21F <#F constraint is a unique identifier for a ro% %it!in a
database table. #"ery table s!ould !a"e a primary ?ey constraint to
uniquely identify eac! ro% and only one primary ?ey constraint can be
created for eac! table. !e primary ?ey constraints are used to
enforce entity integrity.
2 $7&Q$# constraint enforces t!e uniqueness of t!e "alues in a set of
columns, so no duplicate "alues are entered. !e unique ?ey
constraints are used to enforce entity integrity as t!e primary ?ey
constraints.
2 @51#&87 <#F constraint pre"ents any actions t!at %ould destroy
lin? bet%een tables %it! t!e corresponding data "alues. 2 foreign ?ey
in one table points to a primary ?ey in anot!er table. @oreign ?eys
pre"ent actions t!at %ould lea"e ro%s %it! foreign ?ey "alues %!en
t!ere are no primary ?eys %it! t!at "alue. !e foreign ?ey constraints
are used to enforce referential integrity.
2 0G#0< constraint is used to limit t!e "alues t!at can be placed in a
column. !e c!ec? constraints are used to enforce domain integrity.
2 75 7$LL constraint enforces t!at t!e column %ill not accept null
"alues. !e not null constraints are used to enforce domain integrity,
as t!e c!ec? constraints.
Fou can create constraints %!en t!e table is created, as part of t!e
table definition by using t!e 01#2# 2'L# statement.
#/amples
!e follo%ing e/ample creates a c!ec?Jsale 0G#0< constraint on an
employee table-
01#2# 2'L# employee(
#mployee&d &7 75 7$LL,
L7ame >210G21(3K) 75 7$LL,
@7ame >210G21(3K) 75 7$LL,
2ddress >210G21(1KK) 75 7$LL,
Gire.ate .2#&(# 75 7$LL,
Salary (57#F 75 7$LL 057S12&7 c!ec?Jsale 0G#0< (salary L
K)
)
Fou can add constraints to an e/isting table by using t!e 2L#1 2'L#
statement. !e follo%ing e/ample adds a p?Jemployee primary ?ey
constraint on an employee table-
2L#1 2'L# employee
2.. 057S12&7 p?Jemployee 61&(21F <#F (#mployee&d)
Fou can add t!e primary or unique ?ey constraint into an e/isting
table only %!en t!ere are no duplicate ro%s in t!e table. Fou can drop
constraints in an e/isting table by using t!e 2L#1 2'L# statement.
!e follo%ing e/ample drops t!e p?Jemployee primary ?ey constraint
in t!e employee table-
2L#1 2'L# employee
.156 057S12&7 p?Jemployee
Sometimes you need to perform some actions t!at require t!e
@51#&87 <#F or 0G#0< constraints be disabled, for e/ample, your
company do not !ire foreign employees, you made t!e appropriate
constraint, but t!e situation %as c!anged and your boss need to !ire
t!e foreign employee, but only t!is one. &n t!is case, you need to
disable t!e constraint by using t!e 2L#1 2'L# statement. 2fter
t!ese actions %ill be performed, you can re3enable t!e @51#&87 <#F
and 0G#0< constraints by using t!e 2L#1 2'L# statement.
*. What is -rimary .ey?
!e primary ?ey of a relational table uniquely identifies eac! record in
t!e table. &t can eit!er be a normal attribute t!at is guaranteed to be
unique (suc! as Social Security 7umber in a table %it! no more t!an
one record per person) or it can be generated by t!e .'(S (suc! as a
globally unique identifier, or 8$&., in (icrosoft SQL Ser"er). 6rimary
?eys may consist of a single attribute or multiple attributes in
combination.
SQL Ser/er-
!0E1TE T12LE !ustomer
$S3D inte#er -03M104 .E4,
Last5)ame /ar"har$36%,
+irst5)ame /ar"har$36%%7
'elo% are e/amples for specifying a primary ?ey by altering a table-
SQL Ser/er-
1LTE0 T12LE !ustomer 1DD -03M104 .E4 $S3D%7
8. What is +orei#n .ey?
2 foreign ?ey is a field (or fields) t!at points to t!e primary ?ey of
anot!er table. !e purpose of t!e foreign ?ey is to ensure referential
integrity of t!e data. &n ot!er %ords, only "alues t!at are supposed to
appear in t!e database are permitted.
SQL Ser/er-
!0E1TE T12LE 90DE0S
$9rer53D inte#er primary :ey,
9rer5Date atetime,
!ustomer5S3D inte#er referen"es !;ST9ME0$S3D%,
1mount ou<le%7
SQL Ser/er-
1LTE0 T12LE 90DE0S
1DD +90E3=) .E4 $"ustomer5si% 0E+E0E)!ES
!;ST9ME0$S3D%7
).%!at is t!e difference bet%een primary ?ey M foreign ?eyB
primary ?ey is used to identify a ro% and it doesnot allo%
null "alues. &t a"oids duplication of ro%s.
%!ereas foreign ?ey refers to a column or combination of
columns included in t!e defenition of referential integrity.
16. SQL 3nte#rity !onstraints
&ntegrity 0onstraints are used to apply business rules for t!e database
tables.
!e constraints a"ailable in SQL are +orei#n .ey, )ot )ull, ;nique,
!he":.
0onstraints can be defined in t%o %ays
1) !e constraints can be specified immediately after t!e column definition.
!is is called column3le"el definition.
2) !e constraints can be specified after all t!e columns are defined. !is is
called table3le"el definition.
1% SQL -rimary :ey'
!is constraint defines a column or combination of columns %!ic! uniquely
identifies eac! ro% in t!e table.
Synta& to efine a -rimary :ey at "olumn le/el'
column name datatype N057S12&7 constraintJnameO 61&(21F <#F
Synta& to efine a -rimary :ey at ta<le le/el'
N057S12&7 constraintJnameO 61&(21F <#F (columnJname1,columnJname2,..)
column_name1, column_name2 are t!e names of t!e columns %!ic! define t!e
primary <ey.
!e synta/ %it!in t!e brac?et i.e. N057S12&7 constraintJnameO is optional.
+or E&ample' o create an employee table %it! 6rimary <ey constraint, t!e
query %ould be li?e.
-rimary .ey at ta<le le/el'
01#2# 2'L# employee
( id number(,) 61&(21F <#F,
name c!ar(2K),
dept c!ar(1K),
age number(2),
salary number(1K),
location c!ar(1K)
)P
or
01#2# 2'L# employee
( id number(,) 057S12&7 empJidJp? 61&(21F <#F,
name c!ar(2K),
dept c!ar(1K),
age number(2),
salary number(1K),
location c!ar(1K)
)P
-rimary .ey at ta<le le/el'
01#2# 2'L# employee
( id number(,),
name c!ar(2K),
dept c!ar(1K),
age number(2),
salary number(1K),
location c!ar(1K),
057S12&7 empJidJp? 61&(21F <#F (id)
)P
2% SQL +orei#n :ey or 0eferential 3nte#rity '
!is constraint identifies any column referencing t!e 61&(21F <#F in
anot!er table. &t establis!es a relations!ip bet%een t%o columns in t!e
same table or bet%een different tables. @or a column to be defined as a
@oreign <ey, it s!ould be a defined as a 6rimary <ey in t!e table %!ic! it is
referring. 5ne or more columns can be defined as @oreign ?ey.
Synta& to efine a +orei#n :ey at "olumn le/el'
N057S12&7 constraintJnameO 1#@#1#70#S 1eferencedJableJname(columnJname)
Synta& to efine a +orei#n :ey at ta<le le/el'
N057S12&7 constraintJnameO @51#&87 <#F(columnJname) 1#@#1#70#S
referencedJtableJname(columnJname)P
+or E&ample'
1) Lets use t!e QproductQ table and QorderJitemsQ.
+orei#n .ey at "olumn le/el'
01#2# 2'L# product
( productJid number(,) 057S12&7 pdJidJp? 61&(21F <#F,
productJname c!ar(2K),
supplierJname c!ar(2K),
unitJprice number(1K)
)P
01#2# 2'L# orderJitems
( orderJid number(,) 057S12&7 odJidJp? 61&(21F <#F,
productJid number(,) 057S12&7 pdJidJf? 1#@#1#70#S, product(productJid),
productJname c!ar(2K),
supplierJname c!ar(2K),
unitJprice number(1K)
)P
+orei#n .ey at ta<le le/el'
01#2# 2'L# orderJitems
( orderJid number(,) ,
productJid number(,),
productJname c!ar(2K),
supplierJname c!ar(2K),
unitJprice number(1K)
057S12&7 odJidJp? 61&(21F <#F(orderJid),
057S12&7 pdJidJf? @51#&87 <#F(productJid) 1#@#1#70#S product(productJid)
)P
2) &f t!e employee table !as a =mgrJid= i.e, manager id as a foreign ?ey
%!ic! references primary ?ey =id= %it!in t!e same table, t!e query %ould be
li?e,
01#2# 2'L# employee
( id number(,) 61&(21F <#F,
name c!ar(2K),
dept c!ar(1K),
age number(2),
mgrJid number(,) 1#@#1#70#S employee(id),
salary number(1K),
location c!ar(1K)
)P
3% SQL )ot )ull !onstraint '
!is constraint ensures all ro%s in t!e table contain a definite "alue for t!e
column %!ic! is specified as not null. A!ic! means a null "alue is not
allo%ed.
Synta& to efine a )ot )ull "onstraint'
N057S12&7 constraint nameO 75 7$LL
+or E&ample' o create a employee table %it! 7ull "alue, t!e query %ould
be li?e
01#2# 2'L# employee
( id number(,),
name c!ar(2K) 057S12&7 nmJnn 75 7$LL,
dept c!ar(1K),
age number(2),
salary number(1K),
location c!ar(1K)
)P
(% SQL ;nique .ey'
!is constraint ensures t!at a column or a group of columns in eac! ro%
!a"e a distinct "alue. 2 column(s) can !a"e a null "alue but t!e "alues
cannot be duplicated.
Synta& to efine a ;nique :ey at "olumn le/el'
N057S12&7 constraintJnameO $7&Q$#
Synta& to efine a ;nique :ey at ta<le le/el'
N057S12&7 constraintJnameO $7&Q$#(columnJname)
+or E&ample' o create an employee table %it! $nique ?ey, t!e query
%ould be li?e,
;nique .ey at "olumn le/el'
01#2# 2'L# employee
( id number(,) 61&(21F <#F,
name c!ar(2K),
dept c!ar(1K),
age number(2),
salary number(1K),
location c!ar(1K) $7&Q$#
)P
or
01#2# 2'L# employee
( id number(,) 61&(21F <#F,
name c!ar(2K),
dept c!ar(1K),
age number(2),
salary number(1K),
location c!ar(1K) 057S12&7 locJun $7&Q$#
)P
;nique .ey at ta<le le/el'
01#2# 2'L# employee
( id number(,) 61&(21F <#F,
name c!ar(2K),
dept c!ar(1K),
age number(2),
salary number(1K),
location c!ar(1K),
057S12&7 locJun $7&Q$#(location)
)P
>% SQL !he": !onstraint '
!is constraint defines a business rule on a column. 2ll t!e ro%s must
satisfy t!is rule. !e constraint can be applied for a single column or a
group of columns.
Synta& to efine a !he": "onstraint'
N057S12&7 constraintJnameO 0G#0< (condition)
+or E&ample' &n t!e employee table to select t!e gender of a person, t!e
query %ould be li?e
!he": !onstraint at "olumn le/el'
01#2# 2'L# employee
( id number(,) 61&(21F <#F,
name c!ar(2K),
dept c!ar(1K),
age number(2),
gender c!ar(1) 0G#0< (gender in (=(=,=@=)),
salary number(1K),
location c!ar(1K)
)P
!he": !onstraint at ta<le le/el'
01#2# 2'L# employee
( id number(,) 61&(21F <#F,
name c!ar(2K),
dept c!ar(1K),
age number(2),
gender c!ar(1),
salary number(1K),
location c!ar(1K),
057S12&7 genderJc? 0G#0< (gender in (=(=,=@=))
)P
.ifference bet%een .elete and runcate 0ommandB
!e .#L## command is used to remo"e ro%s from a table. 2 AG#1#
clause can be used to only remo"e some ro%s. &f no AG#1# condition is
specified, all ro%s %ill be remo"ed. 2fter performing a .#L## operation you
need to
05((& or 15LL'20< t!e transaction to ma?e t!e c!ange permanent or to
undo it.
1$702# remo"es all ro%s from a table. !e operation cannot be rolled
bac?. 2s suc!, 1$02# is faster and doesn=t use as muc! undo space as a
.#L##.
!e .156 command remo"es a table from t!e database. 2ll t!e tables=
ro%s,
inde/es and pri"ileges %ill also be remo"ed. !e operation cannot be rolled
bac?.
.156 and 1$702# are ..L commands, %!ereas .#L## is a .(L
command. !erefore .#L## operations can be rolled bac? (undone), %!ile
.156 and 1$702# operations cannot be rolled bac?.
11. 7ame fi"e 'uilt in functions in SQLB
)umeri"al +un"tion ' 2'S(d) 205S(d) 2S&7(d) 227(d) 2272(a,b)
'&27.(a,b) '&51(a,b) '&;51(a,b) 0#&L&78(d) 05S(d) 05(d)
.#81##S(d) #;6(d) @L551(d) L58(d) L581K(d) (5.(a,b) 65A#1(a,b)
12.&27S(d)
Strin# +un"tions ' 2S0&&(s) '&JL#78G(str) 0G21(c) 0G21JL#78G(str)
05702(str1,str2) .&@@#1#70#(s1,s2) G#;512A(s1)
&7S#1(s,start,len,s2) L02S#(s) L#@(s,count) L#78G(s)
L502#(searc!,s,NstartO) L1&((s) L5A#1(s) 50#JL#78G(str)
12A5G#;(s1) 65S&&57 (Rstring e/pressionL &7 Rstring e/pression 1L)
1#6#2(s,count) 1#6L20#(s,replace,s2) 1&8G(s,count) 11&((s)
S5$7.#;(s) S620#(count)
Date an Time +un"tions ' .2#.&@@(string, datetime1, datetime2)
.2F72(#(date) .2F5@(57G(date) .2F5@A##<(date) .2F5@F#21(date)
G5$1(time) (&7$#(time) (57G(date) (57G72(#(date) 75A()
S#057.(time)
System +un"tions' .22'2S#() $S#1() 0$11#7J$S#1 &.#7&F()
1##re#ate +un"tion' 2ggregate functions are used to perform some
computation against group of "alues. #ac! field in a table can be t!oug!t of
as one group of "alues. $se of t!e 815$6 'F clause in a select statement
allo%s t!e creation of multiple "alue groups
05$7( 2ny E S ) (&7(d) (2;(d) S$((d) 2>8(d) S5(#(b) #>#1F(b)
12. Data 3nte#rity
#nforcing data integrity ensures t!e quality of t!e data in t!e database. @or
e/ample, if an employee is entered %it! anemployee5i "alue of 123, t!e
database s!ould not allo% anot!er employee to !a"e an &. %it! t!e same "alue. &f
you !a"e an employee5ratin# column intended to !a"e "alues ranging
from 1 to >, t!e database s!ould not accept a "alue of ,. &f t!e table !as
a ept5i column t!at stores t!e department number for t!e employee, t!e
database s!ould allo% only "alues t!at are "alid for t!e department numbers in t!e
company.
%o important steps in planning tables are to identify "alid "alues for a column and
to decide !o% to enforce t!e integrity of t!e data in t!e column. .ata integrity falls
into t!ese categories-
#ntity integrity
.omain integrity
1eferential integrity
$ser3defined integrity
Entity 3nte#rity
#ntity integrity defines a ro% as a unique entity for a particular table. #ntity
integrity enforces t!e integrity of t!e identifier column(s) or t!e primary ?eyof a
table (t!roug! inde/es, $7&Q$# constraints, 61&(21F <#F constraints, or
&.#7&F properties).
Domain 3nte#rity
.omain integrity is t!e "alidity of entries for a gi"en column. Fou can enforce
domain integrity by restricting t!e type (t!roug! data types), t!e format (t!roug!
0G#0< constraints and rules), or t!e range of possible "alues (t!roug! @51#&87
<#F constraints, 0G#0< constraints, .#@2$L definitions, 75 7$LL definitions,
and rules).
0eferential 3nte#rity
1eferential integrity preser"es t!e defined relations!ips bet%een tables %!en
records are entered or deleted. &n (icrosoftT SQL Ser"erU 2KKK, referential
integrity is based on relations!ips bet%een foreign ?eys and primary ?eys or
bet%een foreign ?eys and unique ?eys (t!roug! @51#&87 <#F and 0G#0<
constraints). 1eferential integrity ensures t!at ?ey "alues are consistent across
tables. Suc! consistency requires t!at t!ere be no references to none/istent "alues
and t!at if a ?ey "alue c!anges, all references to it c!ange consistently t!roug!out
t!e database.
13. What is the ifferen"e <et?een primary :ey @ forei#n
:ey?
primary ?ey is used to identify a ro% and it doesnot allo%
null "alues. &t a"oids duplication of ro%s.
%!ereas foreign ?ey refers to a column or combination of
columns included in t!e defenition of referential integrity.
1(. What are the a/anta#es an isa/anta#es of primary :ey an forei#n :ey
in SQL?
-rimary :ey
1/anta#es
1) &t is a unique ?ey on %!ic! all t!e ot!er candidate ?eys are functionally
dependent
Disa/anta#e
1) !ere can be more t!an one ?eys on %!ic! all t!e ot!er attributes are
dependent on.
+orei#n .ey
1/anta#e
1) &t allo%s refrencing anot!er table using t!e primary ?ey for t!e ot!er table
1( . 9perators
2n operator is a symbol specifying an action t!at is performed on one or more
e/pressions. (icrosoftT SQL Ser"erU 2KKK uses t!ese operator categories-
2rit!metic operators
2ssignment operator
'it%ise operators
0omparison operators
Logical operators
String concatenation operator
$nary operators
1> . +i/e mathemati"al fun"tions?
(at!ematical operators are pro"ided for many 6ostgreSQL types. @or types %it!out
common mat!ematical con"entions for all possible permutations (e.g., date9time types) %e
describe t!e actual be!a"ior in subsequent sections.
abs(/)
(same
as /)
absolute "alue abs(31*.+) 1*.+
cbrt(dp) dp cube root cbrt(2*.K) 3
ceil(dp or numeric)
(same as
input)
smallest
integer not
less t!an
argument
ceil(3+2.V) 3+2
degrees(dp) dp
radians to
degrees
degrees(K.,) 2V.W+*VV)*,W,+12
e/p(dp or numeric)
(same as
input)
e/ponential e/p(1.K) 2.*1V2V1V2V+,)K,
floor(dp or numeric)
(same as
input)
largest integer
not greater
t!an
argument
floor(3+2.V) 3+3
ln(dp or numeric)
(same as
input)
natural
logarit!m
ln(2.K) K.W)31+*1VK,,))+,
log(dp or numeric)
(same as
input)
base 1K
logarit!m
log(1KK.K) 2
log(b numeric, /numeric) numeric
logarit!m to
base b
log(2.K, W+.K) W.KKKKKKKKKK
mod(y, /)
(same as
argument
types)
remainder
of y9/
mod(),+) 1
pi() dp Q6iQ constant pi() 3.1+1,)2W,3,V)*)
po%(/ dp, e dp) dp
raise a
number to
e/ponent e
po%().K, 3.K) *2)
po%(/ numeric, enumeric) numeric
raise a
number to
e/ponent e
po%().K, 3.K) *2)
radians(dp) dp
degrees to
radians
radians(+,.K) K.*V,3)V1W33)*++V
random() dp
random "alue
bet%een K.K
and 1.K
random()
round(dp or numeric)
(same as
input)
round to
nearest
integer
round(+2.+) +2
round(" numeric, sinteger) numeric
round
to s decimal
places
round(+2.+3V2,
2)
+2.++
sign(dp or numeric)
(same as
input)
sign of t!e
argument (31,
K, X1)
sign(3V.+) 31
sqrt(dp or numeric)
(same as
input)
square root sqrt(2.K) 1.+1+213,W23*31
trunc(dp or numeric)
(same as
input)
truncate
to%ard zero
trunc(+2.V) +2
trunc(" numeric, sinteger) numeric
truncate
to s decimal
places
trunc(+2.+3V2,
2)
+2.+3
1,. Tri#nomentri" +un"tions?
+un"tion Des"ription
acos(/) in"erse cosine
asin(/) in"erse sine
atan(/) in"erse tangent
atan2(/, y) in"erse tangent of /9y
cos(/) cosine
cot(/) cotangent
sin(/) sine
tan(/) tangent
1A. SQL Strin# +un"tions an 9perators
+un"tion
0etu
rn
Type
Des"riptio
n
E&ample 0esult
string EE string te/t
String
concatenation
=6ost= EE
=greSQL=
6ostgreSQ
L
bitJlengt!(string)
intege
r
7umber of
bits in string
bitJlengt!(=4ose
=)
32
c!arJlengt!(string)
orc!aracterJlengt!(string)
intege
r
7umber of
c!aracters in
string
c!arJlengt!(=4o
se=)
+
con"ert(string usingcon"ersionJna
me)
te/t
0!ange
encoding
using
specified
con"ersion
name.
0on"ersions
can be
defined
by 01#2#
057>#1S&57
. 2lso t!ere
are some pre3
defined
con"ersion
names.
See Ta<le ,B
8 for
a"ailable
con"ersion
names.
con"ert(=6ostgr
eSQL= using
isoJVV,)J1Jto
JutfJV)
=6ostgreS
QL=in
$nicode
($@3V)
encoding
lo%er(string) te/t
0on"ert string
to lo%er case
lo%er(=5(=) tom
octetJlengt!(string)
intege
r
7umber of
bytes in
string
octetJlengt!(=4o
se=)
+
o"erlay(string placingstring from inte
ger NforintegerO)
te/t
&nsert
substring
o"erlay(=////a
s= placing =!om=
from 2 for +)
!omas
position(substring instring)
intege
r
Location of
specified
substring
position(=om= in
=!omas=)
3
substring(string NfromintegerO
Nfor integerO)
te/t
#/tract
substring
substring(=!o
mas= from 2 for
3)
!om
substring(string frompattern) te/t
#/tract
substring
matc!ing
65S&; regular
e/pression
substring(=!o
mas= from =...
Y=)
mas
substring(string frompattern for esc
ape)
te/t #/tract
substring
matc!ing SQL
regular
substring(=!o
mas= from
=Z[QoJa[QJ=
oma
+un"tion
0etu
rn
Type
Des"riptio
n
E&ample 0esult
e/pression for =[=)
trim(Nleading E trailing E bot!O
Nc!aractersO fromstring)
te/t
1emo"e t!e
longest string
containing
only
t!ec!aracte
rs (a space
by default)
from t!e
beginning9en
d9bot! ends
of t!e string
trim(bot! =/=
from =/om//=)
om
upper(string) te/t
0on"ert string
to upper case
upper(=tom=) 5(
18. DateCTime +un"tions
)ame
0eturn
Type
Des"ription E&ample 0esult
age(timestamp) inter"al Subtract from today
age(timestamp
=1),*3KW313=)
+3 years V mons
3 days
age(timestamp,timestamp
)
inter"al Subtract arguments
age(=2KK13K+3
1K=, timestamp
=1),*3KW313=)
+3 years ) mons
2* days
currentJdate date
oday=s dateP
see Se"tion ,.8.(

currentJtime
time %it!
time zone
ime of dayP
see Se"tion ,.8.(

currentJtimestamp
timestam
p %it!
time zone
.ate and timeP
see Se"tion ,.8.(

dateJpart(te/t,timestamp
)
double
precision
8et subfield
(equi"alent
to e/tract)P see
also <elo?
dateJpart(=!our=,
timestamp
=2KK13K231W
2K-3V-+K=)
2K
dateJpart(te/t,inter"al)
double
precision
8et subfield
(equi"alent
to e/tract)P see
also <elo?
dateJpart(=mont
!=, inter"al =2
years 3 mont!s=)
3
dateJtrunc(te/t,timestam
p)
timestam
p
runcate to specified
precisionP see
also Se"tion ,.8.2
dateJtrunc(=!our=
, timestamp
=2KK13K231W
2K-3V-+K=)
2KK13K231W
2K-KK-KKXKK
e/tract(field fromtimesta
mp)
double
precision
8et subfieldP see
also Se"tion ,.8.1
e/tract(!our
from timestamp
=2KK13K231W
2K-3V-+K=)
2K
e/tract(field frominter"al)
double
precision
8et subfieldP see
also Se"tion ,.8.1
e/tract(mont!
from inter"al =2
years 3 mont!s=)
3
isfinite(timestamp) boolean est for finite time isfinite(timestam true
)ame
0eturn
Type
Des"ription E&ample 0esult
stamp (neit!er
in"alid nor infinity)
p =2KK13K231W
21-2V-3K=)
isfinite(inter"al) boolean
est for finite
inter"al
isfinite(inter"al =+
!ours=)
true
localtime time
ime of dayP
see Se"tion ,.8.(

localtimestamp
timestam
p
.ate and timeP
see Se"tion ,.8.(

no%()
timestam
p %it!
time zone
0urrent date and
time (equi"alent
tocurrentJtimestamp
)P see Se"tion
,.8.(

timeofday() te/t
0urrent date and
timeP see Se"tion
,.8.(
timeofday()
Aed @eb 21
1*-K1-13.KKK12
W 2KK1 #S
1D. 1##re#ate +un"tions
+un"tion 1r#ument Type0eturn Type
Des"ripti
on

a"g(e/pres
sion)
smallint, integer,bigi
nt, real, double
precision, numeric,
or inter"al.
numeric for any integer type
argument, double precision for a floating3
point argument, ot!er%ise t!e same as t!e
argument data type
t!e a"erage
(arit!metic
mean) of all
input "alues

count(S) bigint
number of
input "alues

count(e/pre
ssion)
any bigint
number of
input "alues
for %!ic!
t!e "alue
ofe/pressi
on is not
null

ma/(e/pres
sion)
any numeric, string,
or date9time type
same as argument type
ma/imum
"alue
ofe/pressi
on across
all input
"alues

min(e/pres
sion)
any numeric, string,
or date9time type
same as argument type
minimum
"alue
ofe/pressi
on across
all input
"alues

stdde"(e/pr
ession)
smallint, integer,bigi
nt, real, double
precision,
ornumeric.
double precision for floating3point
arguments, ot!er%ise numeric.
sample
standard
de"iation of
t!e input
"alues

sum(e/pres
smallint, integer,bigi bigint for smallint or integer arguments,nu sum
+un"tion 1r#ument Type0eturn Type
Des"ripti
on

sion)
nt, real, double
precision, numeric,
or inter"al
meric for bigint arguments, double
precisionfor floating3point arguments,
ot!er%ise t!e same as t!e argument data
type
of e/pressi
onacross all
input "alues
"ariance(e/p
ression)
smallint, integer,bigi
nt, real, double
precision,
ornumeric.
double precision for floating3point
arguments, ot!er%ise numeric.
sample
"ariance of
t!e input
"alues
(square of
t!e sample
standard
de"iation)
26 ?hat is ha/in# "lause?
SQL' E1F3)= !lause
!e G2>&78 clause is used in combination %it! t!e 815$6 'F clause. &t can be used in a
S#L#0 statement to filter t!e records t!at a 815$6 'F returns.
!e synta/ for t!e G2>&78 clause is-
SELE!T "olumn1, "olumn2, ... "olumn5n, a##re#ate5fun"tion $e&pression%
+09M ta<les
WEE0E prei"ates
=09;- 24 "olumn1, "olumn2, ... "olumn5n
E1F3)= "onition1 ... "onition5n7
aggregateJfunction can be a function suc! as S;M, !9;)T, M3), or M1G.
E&ample usin# the S;M fun"tion
@or e/ample, you could also use t!e S$( function to return t!e name of t!e department
and t!e total sales (in t!e associated department). !e G2>&78 clause %ill filter t!e results
so t!at only departments %it! sales greater t!an Y1KKK %ill be returned.
SELE!T epartment, S;M$sales% as HTotal salesH
+09M orer5etails
=09;- 24 epartment
E1F3)= S;M$sales% I 16667
E&ample usin# the !9;)T fun"tion
@or e/ample, you could use t!e 05$7 function to return t!e name of t!e department and
t!e number of employees (in t!e associated department) t!at ma?e o"er Y2,,KKK 9 year.
!e G2>&78 clause %ill filter t!e results so t!at only departments %it! more t!an 1K
employees %ill be returned.
SELE!T epartment, !9;)T$J% as H)um<er of employeesH
+09M employees
WEE0E salary I 2>666
=09;- 24 epartment
E1F3)= !9;)T$J% I 167
E&ample usin# the M3) fun"tion
@or e/ample, you could also use t!e (&7 function to return t!e name of eac! department
and t!e minimum salary in t!e department. !e G2>&78 clause %ill return only t!ose
departments %!ere t!e starting salary is Y3,,KKK.
SELE!T epartment, M3)$salary% as HLo?est salaryH
+09M employees
=09;- 24 epartment
E1F3)= M3)$salary% K 3>6667
E&ample usin# the M1G fun"tion
@or e/ample, you could also use t!e (2; function to return t!e name of eac! department
and t!e ma/imum salary in t!e department. !e G2>&78 clause %ill return only t!ose
departments %!ose ma/imum salary is less t!an Y,K,KKK.
SELE!T epartment, M1G$salary% as HEi#hest salaryH
+09M employees
=09;- 24 epartment
E1F3)= M1G$salary% L >66667
21. SQL' =09;- 24 !lause
!e 815$6 'F clause can be used in a S#L#0 statement to collect data across multiple
records and group t!e results by one or more columns.
!e synta/ for t!e 815$6 'F clause is-
SELE!T "olumn1, "olumn2, ... "olumn5n, a##re#ate5fun"tion $e&pression%
+09M ta<les
WEE0E prei"ates
=09;- 24 "olumn1, "olumn2, ... "olumn5n7
aggregateJfunction can be a function suc! as S;M, !9;)T, M3), or M1G.
E&ample usin# the S;M fun"tion
@or e/ample, you could also use t!e S$( function to return t!e name of t!e department
and t!e total sales (in t!e associated department).
SELE!T epartment, S;M$sales% as HTotal salesH
+09M orer5etails
=09;- 24 epartment7
'ecause you !a"e listed one column in your S#L#0 statement t!at is not encapsulated in
t!e S$( function, you must use a 815$6 'F clause. !e department field must, t!erefore,
be listed in t!e 815$6 'F section.
E&ample usin# the !9;)T fun"tion
@or e/ample, you could use t!e 05$7 function to return t!e name of t!e department and
t!e number of employees (in t!e associated department) t!at ma?e o"er Y2,,KKK 9 year.
SELE!T epartment, !9;)T$J% as H)um<er of employeesH
+09M employees
WEE0E salary I 2>666
=09;- 24 epartment7
E&ample usin# the M3) fun"tion
@or e/ample, you could also use t!e (&7 function to return t!e name of eac! department
and t!e minimum salary in t!e department.
SELE!T epartment, M3)$salary% as HLo?est salaryH
+09M employees
=09;- 24 epartment7
E&ample usin# the M1G fun"tion
@or e/ample, you could also use t!e (2; function to return t!e name of eac! department
and t!e ma/imum salary in t!e department.
SELE!T epartment, M1G$salary% as HEi#hest salaryH
+09M employees
=09;- 24 epartment7
2,. 'rea? Statements in SQLB
!e brea? command can be used to create report brea?s. A!en "alue
in a column c!anges a brea? e"ent is generated and t!e action is
e/ecuted. !is %ay you can s?ip a line (S<&6 1) %!en a "alue c!anges
in a column. &n addition duplicate "alues can be suppressed (75.$6).
Fou can specify more t!an one brea? column by adding more on
clauses at t!e end. 'rea? Statements %!ic! are
1. 0ompute 2. 0ompute 'y
2W. A!at is :oins in SQLB
SQL M93)
!e :5&7 ?ey%ord is used in an SQL statement to query data from
t%o or more tables, based on a relations!ip bet%een certain columns
in t!ese tables.
ables in a database are often related to eac! ot!er %it! ?eys.
2 primary ?ey is a column (or a combination of columns) %it! a
unique "alue for eac! ro%. #ac! primary ?ey "alue must be unique
%it!in t!e table. !e purpose is to bind data toget!er, across tables,
%it!out repeating all of t!e data in e"ery table.
Loo? at t!e Q6ersonsQ table-
-53 Last)ame +irst)ame 1ress !ity
1 Gansen 5la imotei"n 1K Sandnes
2 S"endson o"e 'org"n 23 Sandnes
3 6ettersen <ari Storgt 2K Sta"anger
7ote t!at t!e Q6J&dQ column is t!e primary ?ey in t!e Q6ersonsQ table.
!is means t!at no t%o ro%s can !a"e t!e same 6J&d. !e 6J&d
distinguis!es t%o persons e"en if t!ey !a"e t!e same name.
7e/t, %e !a"e t!e Q5rdersQ table-
953 9rer)o -53
1 **V), 3
2 ++W*V 3
3 22+,W 1
+ 2+,W2 1
, 3+*W+ 1,
7ote t!at t!e Q5J&dQ column is t!e primary ?ey in t!e Q5rdersQ table
and t!at t!e Q6J&dQ column refers to t!e persons in t!e Q6ersonsQ
table %it!out using t!eir names.
7otice t!at t!e relations!ip bet%een t!e t%o tables abo"e is t!e
Q6J&dQ column.
2A. Different SQL M93)s
'efore %e continue %it! e/amples, %e %ill list t!e types of :5&7 you
can use, and t!e differences bet%een t!em.
M93)- 1eturn ro%s %!en t!ere is at least one matc! in bot!
tables
LE+T M93)- 1eturn all ro%s from t!e left table, e"en if t!ere are
no matc!es in t!e rig!t table
03=ET M93)- 1eturn all ro%s from t!e rig!t table, e"en if t!ere
are no matc!es in t!e left table
+;LL M93)- 1eturn ro%s %!en t!ere is a matc! in one of t!e
tables
2V. SQL 3))E0 M93) .ey?or
!e &77#1 :5&7 ?ey%ord return ro%s %!en t!ere is at least one
matc! in bot! tables.
SQL 3))E0 M93) Synta&
S#L#0 columnJname(s)
@15( tableJname1
&77#1 :5&7 tableJname2
57 tableJname1.columnJname\tableJname2.columnJname
-S' &77#1 :5&7 is t!e same as :5&7.
SQL 3))E0 M93) E&ample
!e Q6ersonsQ table-
-53 Last)ame +irst)ame 1ress !ity
1 Gansen 5la imotei"n 1K Sandnes
2 S"endson o"e 'org"n 23 Sandnes
3 6ettersen <ari Storgt 2K Sta"anger
!e Q5rdersQ table-
953 9rer)o -53
1 **V), 3
2 ++W*V 3
3 22+,W 1
+ 2+,W2 1
, 3+*W+ 1,
7o% %e %ant to list all t!e persons %it! any orders.
Ae use t!e follo%ing S#L#0 statement-
S#L#0 6ersons.Last7ame, 6ersons.@irst7ame, 5rders.5rder7o
@15( 6ersons
&77#1 :5&7 5rders
57 6ersons.6J&d\5rders.6J&d
51.#1 'F 6ersons.Last7ame
!e result3set %ill loo? li?e t!is-
Last)ame +irst)ame 9rer)o
Gansen 5la 22+,W
Gansen 5la 2+,W2
6ettersen <ari **V),
6ettersen <ari ++W*V
!e &77#1 :5&7 ?ey%ord return ro%s %!en t!ere is at least one
matc! in bot! tables. &f t!ere are ro%s in Q6ersonsQ t!at do not !a"e
matc!es in Q5rdersQ, t!ose ro%s %ill 75 be listed.
28. SQL LE+T M93) .ey?or
!e L#@ :5&7 ?ey%ord returns all ro%s from t!e left table
(tableJname1), e"en if t!ere are no matc!es in t!e rig!t table
(tableJname2).
SQL LE+T M93) Synta&
S#L#0 columnJname(s)
@15( tableJname1
L#@ :5&7 tableJname2
57 tableJname1.columnJname\tableJname2.columnJname
-S' &n some databases L#@ :5&7 is called L#@ 5$#1 :5&7.
SQL LE+T M93) E&ample
!e Q6ersonsQ table-
-53 Last)ame +irst)ame 1ress !ity
1 Gansen 5la imotei"n 1K Sandnes
2 S"endson o"e 'org"n 23 Sandnes
3 6ettersen <ari Storgt 2K Sta"anger
!e Q5rdersQ table-
953 9rer)o -53
1 **V), 3
2 ++W*V 3
3 22+,W 1
+ 2+,W2 1
, 3+*W+ 1,
7o% %e %ant to list all t!e persons and t!eir orders 3 if any, from t!e
tables abo"e.
Ae use t!e follo%ing S#L#0 statement-
S#L#0 6ersons.Last7ame, 6ersons.@irst7ame, 5rders.5rder7o
@15( 6ersons
L#@ :5&7 5rders
57 6ersons.6J&d\5rders.6J&d
51.#1 'F 6ersons.Last7ame
!e result3set %ill loo? li?e t!is-
Last)ame +irst)ame 9rer)o
Gansen 5la 22+,W
Gansen 5la 2+,W2
6ettersen <ari **V),
6ettersen <ari ++W*V
S"endson o"e
!e L#@ :5&7 ?ey%ord returns all t!e ro%s from t!e left table
(6ersons), e"en if t!ere are no matc!es in t!e rig!t table (5rders).
SQL 03=ET M93) .ey?or
!e 1&8G :5&7 ?ey%ord 1eturn all ro%s from t!e rig!t table
(tableJname2), e"en if t!ere are no matc!es in t!e left table
(tableJname1).
SQL 03=ET M93) Synta&
S#L#0 columnJname(s)
@15( tableJname1
1&8G :5&7 tableJname2
57 tableJname1.columnJname\tableJname2.columnJname
-S' &n some databases 1&8G :5&7 is called 1&8G 5$#1 :5&7.
SQL 03=ET M93) E&ample
!e Q6ersonsQ table-
-53 Last)ame +irst)ame 1ress !ity
1 Gansen 5la imotei"n 1K Sandnes
2 S"endson o"e 'org"n 23 Sandnes
3 6ettersen <ari Storgt 2K Sta"anger
!e Q5rdersQ table-
953 9rer)o -53
1 **V), 3
2 ++W*V 3
3 22+,W 1
+ 2+,W2 1
, 3+*W+ 1,
7o% %e %ant to list all t!e orders %it! containing persons 3 if any,
from t!e tables abo"e.
Ae use t!e follo%ing S#L#0 statement-
S#L#0 6ersons.Last7ame, 6ersons.@irst7ame, 5rders.5rder7o
@15( 6ersons
1&8G :5&7 5rders
57 6ersons.6J&d\5rders.6J&d
51.#1 'F 6ersons.Last7ame
!e result3set %ill loo? li?e t!is-
Last)ame +irst)ame 9rer)o
Gansen 5la 22+,W
Gansen 5la 2+,W2
6ettersen <ari **V),
6ettersen <ari ++W*V
3+*W+
!e 1&8G :5&7 ?ey%ord returns all t!e ro%s from t!e rig!t table
(5rders), e"en if t!ere are no matc!es in t!e left table (6ersons).
SQL +;LL M93) .ey?or
!e @$LL :5&7 ?ey%ord return ro%s %!en t!ere is a matc! in one of
t!e tables.
SQL +;LL M93) Synta&
S#L#0 columnJname(s)
@15( tableJname1
@$LL :5&7 tableJname2
57 tableJname1.columnJname\tableJname2.columnJname
SQL +;LL M93) E&ample
!e Q6ersonsQ table-
-53 Last)ame +irst)ame 1ress !ity
1 Gansen 5la imotei"n 1K Sandnes
2 S"endson o"e 'org"n 23 Sandnes
3 6ettersen <ari Storgt 2K Sta"anger
!e Q5rdersQ table-
953 9rer)o -53
1 **V), 3
2 ++W*V 3
3 22+,W 1
+ 2+,W2 1
, 3+*W+ 1,
7o% %e %ant to list all t!e persons and t!eir orders, and all t!e orders
%it! t!eir persons.
Ae use t!e follo%ing S#L#0 statement-
S#L#0 6ersons.Last7ame, 6ersons.@irst7ame, 5rders.5rder7o
@15( 6ersons
@$LL :5&7 5rders
57 6ersons.6J&d\5rders.6J&d
51.#1 'F 6ersons.Last7ame
!e result3set %ill loo? li?e t!is-
Last)ame +irst)ame 9rer)o
Gansen 5la 22+,W
Gansen 5la 2+,W2
6ettersen <ari **V),
6ettersen <ari ++W*V
S"endson o"e
3+*W+
!e @$LL :5&7 ?ey%ord returns all t!e ro%s from t!e left table
(6ersons), and all t!e ro%s from t!e rig!t table (5rders). &f t!ere are
ro%s in Q6ersonsQ t!at do not !a"e matc!es in Q5rdersQ, or if t!ere are
ro%s in Q5rdersQ t!at do not !a"e matc!es in Q6ersonsQ, t!ose ro%s
%ill be listed as %ell.
3K. %!at is the difference bet%een &nner 4oin and full outer 4oin in
terms of t!e 7umber of ro%s returnedB
&nner :oin %ill return ro%s t!at !a"e matc!ing records in bot! table,
but in case of full outer 4oin it %ill return all records from left table and
from rig!t table.
&n Left35uter 4oin all t!e record from t!e left3most table %ill be
fetc!ed and if records found in t!e 2nd table t!en field "alues %ill be
displayed else it %ill display 7ull in t!e 2nd table fields.
#/.
0reate table #(672(# (&. int, #(672(# "arc!ar(2K))
01#2# table #(62..1#SS (&. int, 2..1#SS "arc!ar(2K))
&7S#1 &75 #(672(# >2L$#S(1, =.##627=)
&7S#1 &75 #(672(# >2L$#S(2, =12(#SG=)
&7S#1 &75 #(672(# >2L$#S(3, =612.##6=)
&7S#1 &75 #(62..1#SS >2L$#S(1, ='2782L51#=)
&7S#1 &75 #(62..1#SS >2L$#S(2, =.#LG&=)
&7S#1 &75 #(62..1#SS >2L$#S(+, =.#LG&=)
S#L#0 S @15( #(672(#
&. #(672(#
33333333333 33333333333333333333
1 .##627
2 12(#SG
3 612.##6
S#L#0 S @15( #(62..1#SS
&. 2..1#SS
33333333333 33333333333333333333
1 '2782L51#
2 .#LG&
+ .#LG&
#/ample - &77#1 :5&7
\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\
\\\\\\\\\\\\\\\\\\\\\\\\
S#L#0 #(672(#.&., #(672(#.#(672(#, #(62..1#SS.2..1#SS
@15( #(672(#
:5&7 #(62..1#SS 57 #(62..1#SS.&. \ #(672(#.&.
5utput of t!e abo"e :5&7 Query (i.e &nner Query)
33333333333333333333333333333333333333333333333333333
&. #(672(# 2..1#SS
33333333333 33333333333333333333 33333333333333333333
1 .##627 '2782L51#
2 12(#SG .#LG&
#/ample - @$LL 5$#1 :5&7
\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\
\\\\\\\\\\\\\\\\\\\\\\\\
S#L#0 #(672(#.&., #(672(#.#(672(#, #(62..1#SS.2..1#SS
@15( #(672(#
@$LL 5$#1 :5&7 #(62..1#SS 57 #(62..1#SS.&. \ #(672(#.&.
5utput of t!e abo"e @$LL 5$#1 :5&7 Query
33333333333333333333333333333333333333333333333333333
&. #(672(# 2..1#SS
33333333333 33333333333333333333 33333333333333333333
1 .##627 '2782L51#
2 12(#SG .#LG&
3 612.##6 7$LL
7$LL 7$LL .#LG&
#/ample - L#@ 5$#1 :5&7
\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\
\\\\\\\\\\\\\\\\\\\\\\\\
S#L#0 #(672(#.&., #(672(#.#(672(#, #(62..1#SS.2..1#SS
@15( #(672(#
L#@ 5$#1 :5&7 #(62..1#SS 57 #(62..1#SS.&. \ #(672(#.&.
5utput of t!e abo"e L#@ 5$#1 :5&7 Query
33333333333333333333333333333333333333333333333333333
&. #(672(# 2..1#SS
33333333333 33333333333333333333 33333333333333333333
1 .##627 '2782L51#
2 12(#SG .#LG&
3 612.##6 7$LL
#/ample - 1&8G 5$#1 :5&7
\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\
\\\\\\\\\\\\\\\\\\\\\\\\
S#L#0 #(672(#.&., #(672(#.#(672(#, #(62..1#SS.2..1#SS
@15( #(672(#
1&8G 5$#1 :5&7 #(62..1#SS 57 #(62..1#SS.&. \
#(672(#.&.
5utput of t!e abo"e 1&8G 5$#1 :5&7 Query
33333333333333333333333333333333333333333333333333333
&. #(672(# 2..1#SS
33333333333 33333333333333333333 33333333333333333333
1 .##627 '2782L51#
2 12(#SG .#LG&
7$LL 7$LL .#LG&
31 .A!at is t!e difference bet%een left outer 4oin and (X) outer 4oinB
A!en & use left outer 4oin t!e cost of t!e SQL query is !ig!, and %!en
& use t!e (X) operator t!e 4oin cost is muc! less.
!ere are t!ree types of outer 4oins, left, rig!t and full. 2 left outer
4oin contains all records of t!e QleftQ table e"en if it !as no matc!es in
t!e Qrig!tQ table specified in t!e 4oin. 2 rig!t outer 4oin contains all
records in t!e Qrig!tQ able e"en if it !as no matc!es in t!e QleftQ table.
2 full outer 4oin contains all records of bot! t!e left and rig!t tables.
32. SQL Su<query
Subquery or &nner query or 7ested query is a query in a query. 2
subquery is usually added in t!e AG#1# 0lause of t!e sql statement.
(ost of t!e time, a subquery is used %!en you ?no% !o% to searc!
for a "alue using a S#L#0 statement, but do not ?no% t!e e/act
"alue.
Subqueries are an alternate %ay of returning data from multiple
tables.
Subqueries can be used %it! t!e follo%ing sql statements along %it!
t!e comparision operators li?e \, R, L, L\, R\ etc.
+or E&ample'
1) $sually, a subquery s!ould return only one record, but sometimes
it can also return multiple records %!en used %it! operators li?e &7,
75 &7 in t!e %!ere clause. !e query %ould be li?e,
S#L#0 firstJname, lastJname, sub4ect
@15( studentJdetails
AG#1# games 75 &7 (=0ric?et=, =@ootball=)P
!e output %ould be similar to-
first5name last5name su<Ne"t
3333333333333 3333333333333 3333333333
S!e?ar 8o%da 'adminton
6riya 0!andra 0!ess
2) Lets consider t!e studentJdetails table %!ic! %e !a"e used earlier.
&f you ?no% t!e name of t!e students %!o are studying science
sub4ect, you can get t!eir id=s by using t!is query belo%,
S#L#0 id, firstJname
@15( studentJdetails
AG#1# firstJname &7 (=1a!ul=, =Step!en=)P
but, if you do not ?no% t!eir names, t!en to get t!eir id=s you need to
%rite t!e query in t!is manner,
S#L#0 id, firstJname
@15( studentJdetails
AG#1# firstJname &7 (S#L#0 firstJname
@15( studentJdetails
AG#1# sub4ect\ =Science=)P
5utput-
i first5name
33333333 3333333333333
1KK 1a!ul
1K2 Step!en
&n t!e abo"e sql statement, first t!e inner query is processed first and
t!en t!e outer query is processed.
3) Subquery can be used %it! &7S#1 statement to add ro%s of data
from one or more tables to anot!er table. Lets try to group all t!e
students %!o study (at!s in a table =mat!sJgroup=.
&7S#1 &75 mat!sJgroup(id, name)
S#L#0 id, firstJname EE = = EE lastJname
@15( studentJdetails AG#1# sub4ect\ =(at!s=
+) 2 subquery can be used in t!e S#L#0 statement as follo%s. Lets
use t!e product and orderJitems table defined in t!e sqlJ4oins section.
select p.productJname, p.supplierJname, (select orderJid from
orderJitems %!ere productJid \ 1K1) as orderJid from product p
%!ere p.productJid \ 1K1
prou"t5name supplier5name orer5i
333333333333333333 333333333333333333 3333333333
ele"ision 5nida ,1K3
33.!orrelate Su<query
2 query is called correlated subquery %!en bot! t!e inner query and
t!e outer query are interdependent. @or e"ery ro% processed by t!e
inner query, t!e outer query is processed as %ell. !e inner query
depends on t!e outer query before it can be processed.
S#L#0 p.productJname @15( product p
AG#1# p.productJid \ (S#L#0 o.productJid @15( orderJitems o
AG#1# o.productJid \ p.productJid)P
3(. SQL !omparison .ey?ors
!ere are ot!er comparison ?ey%ords a"ailable in sql %!ic! are used
to en!ance t!e searc! capabilities of a sql query. !ey are Q&7Q,
Q'#A##7...27.Q, Q&S 7$LLQ, QL&<#Q.
Comparision
Operators
Description
L&<# column "alue is similar to specified c!aracter(s).
&7
column "alue is equal to any one of a specified set
of "alues.
'#A##7...27.
column "alue is bet%een t%o "alues, including t!e
end "alues specified in t!e range.
&S 7$LL column "alue does not e/ist.
SQL L3.E 9perator
!e L&<# operator is used to list all ro%s in a table %!ose column
"alues matc! a specified pattern. &t is useful %!en you %ant to searc!
ro%s to matc! a specific pattern, or %!en you do not ?no% t!e entire
"alue. @or t!is purpose %e use a %ildcard c!aracter =Z=.
+or e&ample' o select all t!e students %!ose name begins %it! =S=
S#L#0 firstJname, lastJname
@15( studentJdetails
AG#1# firstJname L&<# =SZ=P
!e output %ould be similar to-
first5name last5name
3333333333333 3333333333333
Step!en @leming
S!e?ar 8o%da
!e abo"e select statement searc!es for all t!e ro%s %!ere t!e first
letter of t!e column firstJname is =S= and rest of t!e letters in t!e
name can be any c!aracter.
!ere is anot!er %ildcard c!aracter you can use %it! L&<# operator. &t
is t!e underscore c!aracter, = J = . &n a searc! string, t!e underscore
signifies a single c!aracter.
+or e&ample' to display all t!e names %it! =a= second c!aracter,
S#L#0 firstJname, lastJname
@15( studentJdetails
AG#1# firstJname L&<# =JaZ=P
!e output %ould be similar to-
first5name last5name
3333333333333 3333333333333
1a!ul S!arma
)9TE'#ac! underscore act as a place!older for only one c!aracter. So
you can use more t!an one underscore. #g- = JJiZ =3t!is !as t%o
underscores to%ards t!e left, =SJJ4Z= 3 t!is !as t%o underscores
bet%een c!aracter =S= and =i=.
SQL 2ETWEE) ... 1)D 9perator
!e operator '#A##7 and 27., are used to compare data for a
range of "alues.
+or E&ample' to find t!e names of t!e students bet%een age 1K to
1, years, t!e query %ould be li?e,
S#L#0 firstJname, lastJname, age
@15( studentJdetails
AG#1# age '#A##7 1K 27. 1,P
!e output %ould be similar to-
first5name last5name a#e
3333333333333 3333333333333
3333
33
1a!ul S!arma 1K
2na4ali '!ag%at 12
S!e?ar 8o%da 1,
SQL 3) 9perator'
!e &7 operator is used %!en you %ant to compare a column %it!
more t!an one "alue. &t is similar to an 51 condition.
+or e&ample' &f you %ant to find t!e names of students %!o are
studying eit!er (at!s or Science, t!e query %ould be li?e,
S#L#0 firstJname, lastJname, sub4ect
@15( studentJdetails
AG#1# sub4ect &7 (=(at!s=, =Science=)P
!e output %ould be similar to-
first5name last5name su<Ne"t
3333333333333 3333333333333
333333333
3
2na4ali '!ag%at (at!s
S!e?ar 8o%da (at!s
1a!ul S!arma Science
Step!en @leming Science
Fou can include more sub4ects in t!e list li?e
(=mat!s=,=science=,=!istory=)
)9TE'!e data used to compare is case sensiti"e.
SQL 3S );LL 9perator
2 column "alue is 7$LL if it does not e/ist. !e &S 7$LL operator is
used to display all t!e ro%s for columns t!at do not !a"e a "alue.
+or E&ample' &f you %ant to find t!e names of students %!o do not
participate in any games, t!e query %ould be as gi"en belo%
S#L#0 firstJname, lastJname
@15( studentJdetails
AG#1# games &S 7$LL
!ere %ould be no output as %e !a"e e"ery student participate in a
game in t!e table studentJdetails, else t!e names of t!e students %!o
do not participate in any games %ould be displayed.
3>. 1LL
!e first of t!e operators t!at %e %ill consider is 2LL. $sed %it! one of
t!e comparison operators, 2LL compares a single "alue against a set
of data from a query. #ac! "alue from t!e query=s results is combined
%it! t!e scalar "alue to generate a single scalar e/pression. &f all of
t!e scalar e/pressions e"aluate to true t!en t!e result of t!e 2LL
e/pression %ill be true. 5t!er%ise t!e result %ill be false, or un?no%n
if t!e results of t!e e/pressions cannot be determined.
o demonstrate %e can e/ecute a query against t!e #ngineers table in
t!e :o'S database. Let=s assume t!at a business rule is to be
implemented t!at specifies t!at no engineer may !a"e an o"ertime
rate lo%er t!an anot!er engineer=s basic !ourly pay. Ae can use 2LL
%it! 75 to ma?e t!is comparison as follo%s-
S#L#0
#ngineer&d,
#ngineer7ame,
5"ertime1ate
@15(
#ngineers
AG#1#
75 5"ertime1ate L\ 2LL (S#L#0 Gourly1ate @15( #ngineers)
#/ecuting t!is query re"eals t!at t!ere are t%o engineers %!ose
o"ertime rate is lo%er t!an permitted.
S9ME an 1)4
!e S5(# and 27F operators pro"ide equi"alent functionality. 2s %it!
t!e 2LL operator, eac! creates a series of scalar e/pressions based
upon a single "alue, a comparison operator and t!e results of a single3
column query. !e difference is t!at t!e S5(# and 27F operators
return true if at least one of t!e generated e/pressions e"aluates as
true.
!ese operators allo% us to modify t!e pre"ious e/ample query so
t!at t!e 75 is no longer required. &nstead, %e can as? for t!e list of
engineers %!o !a"e an o"ertime rate t!at is lo%er t!an any ot!er
engineer=s basic !ourly rate, as follo%s-
S#L#0
#ngineer&d,
#ngineer7ame,
5"ertime1ate
@15(
#ngineers
AG#1#
5"ertime1ate R 27F (S#L#0 Gourly1ate @15( #ngineers)
2s t!e 27F and S5(# operators are equi"alent, you can interc!ange
t!e t%o ?ey%ords %it!out affecting t!e results-
S#L#0
#ngineer&d,
#ngineer7ame,
5"ertime1ate
@15(
#ngineers
AG#1#
5"ertime1ate R S5(# (S#L#0 Gourly1ate @15( #ngineers)
9perator -re"een"e
Ae can no% add all of t!e logical operators, including t!ose seen in
pre"ious articles, to t!e table of precedence.
;nary 2it?ise 9perators
]
1rithmeti" 9perators
S 9 Z
;nary 1rithmeti" 9perators
X 3
1rithmeti" C !on"atenation 9perators
NX arit!meticO NX concatenation)O 3
2it?ise 9perators
M ^ E
!omparison 9perators
\, L, R, L\, R\, RL, _\, _L, _R
Lo#i"al
75 27. 2LL 27F '#A##7 &7 L&<# 51 S5(#
3W. %!at is e/ist in SQL ser"erB
(any times you=re required to %rite query to determine if a record
e/ists. ypically you use t!is to determine %!et!er to insert or update
a records. $sing t!e #;&SS ?ey%ord is a great %ay to accomplis!
t!is.
Gere=s a simple e/ample from t!e pubs database using #;&SS-
if #;&SS (select S
from aut!ors
%!ere auJid \ =1*2332311*W=)
6rint =1ecord e/its 3 $pdate=
#LS#
6rint =1ecord doesn==t e/ist 3 &nsert=
!e #;&SS function ta?es one parameter %!ic! is a SQL statement. &f
any records e/ist t!at matc! t!e criteria it returns true, ot!er%ise it
returns false. !is gi"es you a clean, efficient %ay to %rite a stored
procedure t!at does eit!er an insert or update.
!e ot!er benefit of #;&SS is t!at once it finds a single record t!at
matc!es it stops processing. !is doesn=t !a"e a !uge impact if you=re
c!ec?ing on a primary ?ey. &t does !a"e a big impact if you=re
c!ec?ing for e/istance based on anot!er field. 0onsider t!e follo%ing
t%o queries-
if e/ists (select S
from aut!ors
%!ere state \ =ca=)
6rint =1ecord e/its=
#LS#
6rint =1ecord doesn==t e/ist=
if (select count(S)
from aut!ors
%!ere state \ =1*2332311*W=) L K
6rint =1ecord e/its=
#LS#
6rint =1ecord doesn==t e/ist=
&n t!e pubs database t!ere are only 23 records in t!e aut!ors table.
#"en %it! t!at small number of records, t!e &@ #;&SS "ersion runs +
times faster t!an selecting a count. !is is because it stops as soon as
it finds a single record t!at matc!es t!e criteria. !e second
statement must process all t!e ro%s t!at matc!.
3*. %!at is >ie% in SQL ser"erB
Fie?
2 "ie% is a "irtual table t!at consists of columns from one or more
tables. !oug! it is similar to a table, it is stored in t!e database. &t is
a query stored as an ob4ect. Gence, a "ie% is an ob4ect t!at deri"es its
data from one or more tables. !ese tables are referred to as base or
underlying tables.
5nce you !a"e defined a "ie%, you can reference it li?e any ot!er
table in a database.
2 "ie% ser"es as a security mec!anism. !is ensures t!at users are
able to retrie"e and modify only t!e data seen by t!em. $sers cannot
see or access t!e remaining data in t!e underlying tables. 2 "ie% also
ser"es as a mec!anism to simplify query e/ecution. 0omple/ queries
can be stored in t!e form as a "ie%, and data from t!e "ie% can be
e/tracted using simple queries.
E&ample
0onsider t!e 6ublis!ers table belo%. &f you %ant users to see only t%o
columns in t!e table, you can create a "ie% called "%6ublis!ers t!at
%ill refer to t!e 6ublis!ers table and t!e t%o columns required. Fou
can grant 6ermissions to users to use t!e "ie% and re"o?e 6ermissions
from t!e base 6ublis!ers table. !is %ay, users %ill be able to "ie%
only t!e t%o columns referred to by t!e "ie%. !ey %ill not be able to
query on t!e 6ublis!ers table.
-u<lishers
6ubld 6ub7ame 0ity State 0ountry
K*3W 7e% (oon 'oo?s 'oston (2 $S2
KV** 'innet M Gardly Aas!ington .0 $S2
13V) 2lgodata &nfosystems 'er?eley 02 $S2
1W22 @i"e La?es 6ublis!ing 0!icago &L $S2
FW -u<lishers
6ubld 6ub7ame
K*3W 7e% (oon 'oo?s
KV** 'innet M Gardly
13V) 2lgodata &nfosystems
1W22 @i"e La?es 6ublis!ing
>ie%s ensure t!e security of data by restricting access to t!e follo%ing
data-
Specific ro%s of t!e tables.
Specific columns of t!e tables.
Specific ro%s and columns of t!e tables.
1o%s fetc!ed by using 4oins.
Statistical summary of data in a gi"en tables.
Subsets of anot!er "ie% or a subset of "ie%s and tables.
Some common e/amples of "ie%s are-
2 subset of ro%s or columns of a base table.
2 union of t%o or more tables.
2 4oin of t%o or more tables.
2 statistical summary of base tables.
2 subset of anot!er "ie%, or some combination of "ie%s and
base table.
!reatin# Fie?s
2 "ie% can be created by using t!e 01#2# >&#A statement.
Synta&
01#2# >&#A "ie%Jname
N(columnJnameN,columnJnameO`.)O
NA&G #701F6&57O
2S selectJstatement NA&G 0G#0< 56&57O
A!ere-
"ie%Jname specifies t!e name of t!e "ie% and must follo% t!e rules
for identifiers.
columnJname specifies t!e name of t!e column to be used in "ie%. &f
t!e columnJname option is not specified, t!en t!e "ie% is created %it!
t!e same columns as specified in t!e selectJstatement.
A&G #701F6&57 encrypts t!e te/t for t!e "ie% in t!e syscomments
table.
2S specifies t!e actions t!at %ill be performed by t!e "ie%.
selectJstatement specifies t!e S#L#0 Statement t!at defines a "ie%.
!e "ie% may use t!e data contained in ot!er "ie%s and tables.
A&G 0G#0< 56&57 forces t!e data modification statements to fulfill
t!e criteria gi"en in t!e S#L#0 statement defining t!e "ie%. &t also
ensures t!at t!e data is "isible after t!e modifications are made
permanent.
!e restrictions imposed on "ie%s are as follo%s-
2 "ie% can be created only in t!e current database.
!e name of a "ie% must follo% t!e rules for identifiers and must
not be t!e same as t!at of t!e base table.
2 "ie% can be created only if t!ere is a S#L#0 permission on its
base table.
2 S#L#0 &75 statement cannot be used in "ie% declaration
statement.
2 trigger or an inde/ cannot be defined on a "ie%.
!e 01#2# >&#A statement cannot be combined %it! ot!er SQL
statements in a single batc!.
E&ample
01#2# >&#A "%0ustomer
2S
S#L#0 0ustomer&d, 0ompany 7ame, 6!one
@15( 0ustomers
0reates a "ie% called "%0ustomer. 7ote t!at t!e "ie% is a query
stored as an ob4ect. !e data is deri"ed from t!e columns of t!e base
table 0ustomers.
Fou use t!e "ie% by querying t!e "ie% li?e a table.
S#L#0 S@15( "%0$S5(#1
!e output of t!e S#L#0 statement is-
0ustomer&d 0ompany 7ame 6!one
2L@<& 2lfreds @utter?iste K3K3KK*+321
2757 2ntonio (oreno aqueria (,),,,33)32
()1 ro%s affected)
2 "ie% is a "irtual table.
!is c!apter s!o%s !o% to create, update, and delete a "ie%.
SQL !0E1TE F3EW Statement
&n SQL, a "ie% is a "irtual table based on t!e result3set of an SQL
statement.
2 "ie% contains ro%s and columns, 4ust li?e a real table. !e fields in
a "ie% are fields from one or more real tables in t!e database.
Fou can add SQL functions, AG#1#, and :5&7 statements to a "ie%
and present t!e data as if t!e data %ere coming from one single table.
SQL !0E1TE F3EW Synta&
01#2# >&#A "ie%Jname 2S
S#L#0 columnJname(s)
@15( tableJname
AG#1# condition
)ote' 2 "ie% al%ays s!o%s up3to3date data_ !e database engine
recreates t!e data, using t!e "ie%=s SQL statement, e"ery time a user
queries a "ie%.
SQL !0E1TE F3EW E&amples
&f you !a"e t!e 7ort!%ind database you can see t!at it !as se"eral
"ie%s installed by default.
!e "ie% Q0urrent 6roduct ListQ lists all acti"e products (products t!at
are not discontinued) from t!e Q6roductsQ table. !e "ie% is created
%it! t!e follo%ing SQL-
01#2# >&#A N0urrent 6roduct ListO 2S
S#L#0 6roduct&.,6roduct7ame
@15( 6roducts
AG#1# .iscontinued\7o
Ae can query t!e "ie% abo"e as follo%s-
S#L#0 S @15( N0urrent 6roduct ListO
2not!er "ie% in t!e 7ort!%ind sample database selects e"ery product
in t!e Q6roductsQ table %it! a unit price !ig!er t!an t!e a"erage unit
price-
01#2# >&#A N6roducts 2bo"e 2"erage 6riceO 2S
S#L#0 6roduct7ame,$nit6rice
@15( 6roducts
AG#1# $nit6riceL(S#L#0 2>8($nit6rice) @15( 6roducts)
Ae can query t!e "ie% abo"e as follo%s-
S#L#0 S @15( N6roducts 2bo"e 2"erage 6riceO
2not!er "ie% in t!e 7ort!%ind database calculates t!e total sale for
eac! category in 1))*. 7ote t!at t!is "ie% selects its data from
anot!er "ie% called Q6roduct Sales for 1))*Q-
01#2# >&#A N0ategory Sales @or 1))*O 2S
S#L#0 .&S&70 0ategory7ame,Sum(6roductSales) 2S
0ategorySales
@15( N6roduct Sales for 1))*O
815$6 'F 0ategory7ame
Ae can query t!e "ie% abo"e as follo%s-
S#L#0 S @15( N0ategory Sales @or 1))*O
Ae can also add a condition to t!e query. 7o% %e %ant to see t!e
total sale only for t!e category Q'e"eragesQ-
S#L#0 S @15( N0ategory Sales @or 1))*O
AG#1# 0ategory7ame\='e"erages=
SQL ;patin# a Fie?
Fou can update a "ie% by using t!e follo%ing synta/-
SQL !0E1TE 90 0E-L1!E F3EW Synta&
01#2# 51 1#6L20# >&#A "ie%Jname 2S
S#L#0 columnJname(s)
@15( tableJname
AG#1# condition
7o% %e %ant to add t!e Q0ategoryQ column to t!e Q0urrent 6roduct
ListQ "ie%. Ae %ill update t!e "ie% %it! t!e follo%ing SQL-
01#2# >&#A N0urrent 6roduct ListO 2S
S#L#0 6roduct&.,6roduct7ame,0ategory
@15( 6roducts
AG#1# .iscontinued\7o
SQL Droppin# a Fie?
Fou can delete a "ie% %it! t!e .156 >&#A command.
SQL D09- F3EW Synta&
.156 >&#A "ie%Jname
3V. A!at are t!e differences bet%een tables and "ie%sB
2 table is %!ere you store your data. !e table actually occupies
space on dis?. 2 sample table mig!t be t!e #(6L5F## table %!ic! !as
t!e columns #(6&., #(6J72(#, and S2L21F.
2 "ie% is a stored query. 2 sample "ie% mig!t loo? li?e t!e follo%ing-
01#2# >&#A empJ"ie% 2S S#L#0 empid,empJname @15(
employeeP
7otice t!at in t!e "ie%, & !a"e omitted t!e S2L21F column. Fou can
t!en query t!e "ie% similar to t!e follo%ing-
S#L#0 S @15( empJ"ie%P
5racle %ill ?no% t!is is a "ie% and automatically re%rite t!e query to
return only t!e #(6&. and #(6J72(# columns from t!e #(6L5F##
table.
>ie%s are used primarily to store a common query in t!e database.
Ait!out t!e "ie%, you mig!t !a"e t!e same comple/ query stored in
multiple locations in your application code. &f you need to ma?e a
c!ange, you %ould !a"e to c!ange t!e query in all locations. Go%e"er,
if t!at query %ere in a "ie%, you %ould only !a"e to c!ange it in one
location.
!e ot!er common reason for a "ie% is for security purposes. &f &
grant a user t!e ability to query #(6J>&#A and t!ey do not !a"e
permissions to query t!e #(6L5F## table, t!en t!ey could ne"er see
t!e S2L21F column.
3D. SQL 3ne&
&nde/ in sql is created on e/isting tables to retrie"e t!e ro%s quic?ly.
A!en t!ere are t!ousands of records in a table, retrie"ing information
%ill ta?e a long time. !erefore inde/es are created on columns %!ic!
are accessed frequently, so t!at t!e information can be retrie"ed
quic?ly. &nde/es can be created on a single column or a group of
columns. A!en a inde/ is created, it first sorts t!e data and t!en it
assigns a 15A&. for eac! ro%.
Synta& to "reate 3ne&'
01#2# &7.#; inde/Jname
57 tableJname (columnJname1,columnJname2...)P
Synta& to "reate SQL unique 3ne&'
01#2# $7&Q$# &7.#; inde/Jname
57 tableJname (columnJname1,columnJname2...)P
index_name is t!e name of t!e &7.#;.
table_name is t!e name of t!e table to %!ic! t!e inde/ed
column belongs.
column_name1, column_name2.. is t!e list of columns %!ic!
ma?e up t!e &7.#;.
&n 5racle t!ere are t%o types of SQL inde/ namely, implicit and
e/plicit.
3mpli"it 3ne&es'
!ey are created %!en a column is e/plicity defined %it! 61&(21F
<#F, $7&Q$# <#F 0onstraint.
E&pli"it 3ne&es'
!ey are created using t!e Qcreate inde/.. Q synta/.
)9TE'
1% #"en t!oug! sql inde/es are created to access t!e ro%s in t!e table
quic?ly, t!ey slo% do%n .(L operations li?e &7S#1, $6.2#, .#L##
on t!e table, because t!e inde/es and tables bot! are updated along
%!en a .(L operation is performed. So use inde/es only on columns
%!ic! are used to searc! t!e table frequently.
2% &s is not required to create inde/es on table %!ic! !a"e less data.
3% &n oracle database you can define up to si/teen (1W) columns in an
&7.#;.
+K. Whar is an ine&? What are the types of ine&es? Eo?
many "lustere ine&es "an <e "reate on a ta<le? 3 "reate a
separate ine& on ea"h "olumn of a ta<le. ?hat are the
a/anta#es an isa/anta#es of this approa"h?
&nde/es in SQL Ser"er are similar to t!e inde/es in boo?s. !ey !elp
SQL Ser"er retrie"e t!e data quic?er.
3ne&es are of t?o types. 0lustered inde/es and non3clustered
inde/es. A!en you craete a clustered inde/ on a table, all t!e ro%s in
t!e table are stored in t!e order of t!e clustered inde/ ?ey. So, t!ere
can be only one clustered inde/ per table. 7on3clustered inde/es !a"e
t!eir o%n storage separate from t!e table data storage. 7on3clustered
inde/es are stored as '3tree structures (so do clustered inde/es), %it!
t!e leaf le"el nodes !a"ing t!e inde/ ?ey and it=s ro% locater. !e ro%
located could be t!e 1&. or t!e 0lustered inde/ ?ey, depending up on
t!e absence or presence of clustered inde/ on t!e table.
&f you create an inde/ on eac! column of a table, it impro"es t!e
query performance, as t!e query optimizer can c!oose from all t!e
e/isting inde/es to come up %it! an efficient e/ecution plan. 2t t!e
same t ime, data modification operations (suc! as &7S#1, $6.2#,
.#L##) %ill become slo%, as e"ery time data c!anges in t!e table, all
t!e inde/es need to be updated. 2not!er disad"antage is t!at, inde/es
need dis? space, t!e more inde/es you !a"e, more dis? space is used.
+1Why is a ;)39) 1LL faster than a ;)39)?
$7&57 2LL faster t!an a $7&57 because for union operation ser"er
needs to remo"e t!e duplicate "alues but for union all its not. !ats
%!y t!e $75&7 2LL is fater t!an $7&57 5peration. &t is recommended
t!at if you ?no% t!at t!e union set operation ne"er returns duplicate
"alues t!an you must use $7&57 2LL instead of $7&57.
Eo? many types of ata moels are there?
!ere are no standards in t!is area. 2ut!ors and t!eorists ma?e it up
as t!ey go. !e entity3relations!ip model (#1) !as !undreds of
deri"iti"es (bac!man, c!en, ibm, &.#@1/ etc.). t!e most popular of
t!e 55 models is $nified (odeling Language ($(L). 2ctually $(L and
&.#@1/ are closest to becoming a standard t!at can support soft%are
products. 1ational already !as products and &.#@1/ is t!e language of
#1%in.
.on=t be fooled by t!ese "ariations. !ey all represent t!e same
t!ings, you !a"e to be "ery careful t!at you understand all of t!e non3
standard symbols or you %ill surely ma?e mista?es in interpreting
%!at t!e pictures mean.
What is enormali*ation an ?hen ?oul you #o for it?
2s t!e name indicates, denormalization is t!e re"erse process of
normalization. &t=s t!e controlled introduction of redundancy in to t!e
database design. &t !elps impro"e t!e query performance as t!e
number of 4oins could be reduced.
WhatOs the ifferen"e <et?een a primary :ey an a
unique :ey?
'ot! primary ?ey and unique enforce uniqueness of t!e column on
%!ic! t!ey are defined. 'ut by default primary ?ey creates a clustered
inde/ on t!e column, %!ere are unique creates a nonclustered inde/
by default. 2not!er ma4or difference is t!at, primary ?ey doesn=t allo%
7$LLs, but unique ?ey allo%s one 7$LL only.
Define "aniate :ey, alternate :ey, "omposite :ey.
2 candidate ?ey is one t!at can identify eac! ro% of a table uniquely.
8enerally a candidate ?ey becomes t!e primary ?ey of t!e table. &f
t!e table !as more t!an one candidate ?ey, one of t!em %ill become
t!e primary ?ey, and t!e rest are called alternate ?eys.
2 ?ey formed by combining at least t%o or more columns is called
composite ?ey.
What are efaults? 3s there a "olumn to ?hi"h a efault "anOt
<e <oun?
2 default is a "alue t!at %ill be used by a column, if no "alue is
supplied to t!at column %!ile inserting data. &.#7&F columns and
timestamp columns can=t !a"e defaults bound to t!em. See 01#2#
.#@$2L in boo?s online.
Whar is an ine&? What are the types of ine&es? Eo? many
"lustere ine&es "an <e "reate on a ta<le? 3 "reate a
separate ine& on ea"h "olumn of a ta<le. ?hat are the
a/anta#es an isa/anta#es of this approa"h?
&nde/es in SQL Ser"er are similar to t!e inde/es in boo?s. !ey !elp
SQL Ser"er retrie"e t!e data quic?er.
3ne&es are of t?o types. 0lustered inde/es and non3clustered
inde/es. A!en you craete a clustered inde/ on a table, all t!e ro%s in
t!e table are stored in t!e order of t!e clustered inde/ ?ey. So, t!ere
can be only one clustered inde/ per table. 7on3clustered inde/es !a"e
t!eir o%n storage separate from t!e table data storage. 7on3clustered
inde/es are stored as '3tree structures (so do clustered inde/es), %it!
t!e leaf le"el nodes !a"ing t!e inde/ ?ey and it=s ro% locater. !e ro%
located could be t!e 1&. or t!e 0lustered inde/ ?ey, depending up on
t!e absence or presence of clustered inde/ on t!e table.
&f you create an inde/ on eac! column of a table, it impro"es t!e
query performance, as t!e query optimizer can c!oose from all t!e
e/isting inde/es to come up %it! an efficient e/ecution plan. 2t t!e
same t ime, data modification operations (suc! as &7S#1, $6.2#,
.#L##) %ill become slo%, as e"ery time data c!anges in t!e table, all
t!e inde/es need to be updated. 2not!er disad"antage is t!at, inde/es
need dis? space, t!e more inde/es you !a"e, more dis? space is used.
What are "ursors? E&plain ifferent types of "ursors. What are
the isa/anta#es of "ursors? Eo? "an you a/oi "ursors?
0ursors allo% ro%3by3ro% prcessing of t!e resultsets.
ypes of cursors- Stati", Dynami", +or?arBonly, .eysetBri/en.
See boo?s online for more information.
Disa/anta#es of "ursors' #ac! time you fetc! a ro% from t!e
cursor, it results in a net%or? roundtrip, %!ere as a normal S#L#0
query ma?es only one ro%undtrip, !o%e"er large t!e resultset is.
0ursors are also costly because t!ey require more resources and
temporary storage (results in more &5 operations). @urt!ere, t!ere are
restrictions on t!e S#L#0 statements t!at can be used %it! some
types of cursors.
(ost of t!e times, set based operations can be used instead of
cursors. Gere is an e/ample-
&f you !a"e to gi"e a flat !i?e to your employees using t!e follo%ing
criteria-
Salary bet%een 3KKKK and +KKKK 33 ,KKK !i?e
Salary bet%een +KKKK and ,,KKK 33 *KKK !i?e
Salary bet%een ,,KKK and W,KKK 33 )KKK !i?e
&n t!is situation many de"elopers tend to use a cursor, determine eac!
employee=s salary and update !is salary according to t!e abo"e
formula. 'ut t!e same can be ac!ie"ed by multiple update statements
or can be combined in a single $6.2# statement as s!o%n belo%-
$6.2# tblJemp S# salary \
02S# AG#7 salary '#A##7 3KKKK 27. +KKKK G#7 salary X ,KKK
AG#7 salary '#A##7 +KKKK 27. ,,KKK G#7 salary X *KKK
AG#7 salary '#A##7 ,,KKK 27. W,KKK G#7 salary X 1KKKK
#7.

2not!er situation in %!ic! de"elopers tend to use cursors- Fou need to
call a stored procedure %!en a column in a particular ro% meets
certain condition. Fou don=t !a"e to use cursors for t!is. !is can be
ac!ie"ed using AG&L# loop, as long as t!ere is a unique ?ey to
identify eac! ro%. @or e/amples of using AG&L# loop for ro% by ro%
processing,
What is a Noin an e&plain ifferent types of Noins?
:oins are used in queries to e/plain !o% different tables are related.
:oins also let you select data from a table depending upon data from
anot!er table.
ypes of 4oins- &77#1 :5&7s, 5$#1 :5&7s, 015SS :5&7s. 5$#1
:5&7s are furt!er classified as L#@ 5$#1 :5&7S, 1&8G 5$#1
:5&7S and @$LL 5$#1 :5&7S.
What is a Store -ro"eure?
&ts not!ing but a set of 3SQL statements combined to perform a
single tas? of se"eral tas?s. &ts basically li?e a (acro so %!en you
in"o?e t!e Stored procedure, you actually run a set of statements.
What is the <asi" ifferen"e <et?een "lustere an a nonB
"lustere ine&?
!e difference is t!at, 0lustered inde/ is unique for any gi"en table
and %e can !a"e only one clustered inde/ on a table. !e leaf le"el of
a clustered inde/ is t!e actual data and t!e data is resorted in case of
clustered inde/. A!ereas in case of non3clustered inde/ t!e leaf le"el
is actually a pointer to t!e data in ro%s so %e can !a"e as many non3
clustered inde/es as %e can on t!e db.
What are "ursors?
Aell cursors !elp us to do an operation on a set of data t!at %e
retrei"e by commands suc! as Select columns from table. @or
e/ample - &f %e !a"e duplicate records in a table %e can remo"e it by
declaring a cursor %!ic! %ould c!ec? t!e records during retrei"al one
by one and remo"e ro%s %!ic! !a"e duplicate "alues.
Whi"h T!-C3- port oes SQL Ser/er run on?
SQL Ser"er runs on port 1+33 but %e can also c!ange it for better
security.
!an ?e use Trun"ate "omman on a ta<le ?hi"h is referen"e
<y +90E3=) .E4?
7o. Ae cannot use runcate command on a table %it! @oreign <ey
because of referential integrity.
What is the use of D2!! "ommans?
.'00 stands for database consistency c!ec?er. Ae use t!ese
commands to c!ec? t!e consistency of t!e databases, i.e.,
maintenance, "alidation tas? and status c!ec?s.
What is the ifferen"e <et?een a E1F3)= !L1;SE an a
WEE0E !L1;SE?
Ga"ing 0lause is basically used only %it! t!e 815$6 'F function in a
query. AG#1# 0lause is applied to eac! ro% before t!ey are part of
t!e 815$6 'F function in a query.
What is a Lin:e Ser/er?
Lin?ed Ser"ers is a concept in SQL Ser"er by %!ic! %e can add ot!er
SQL Ser"er to a 8roup and query bot! t!e SQL Ser"er dbs using 3
SQL Statements.
!an you lin: only other SQL Ser/ers or any ata<ase ser/ers
su"h as 9ra"le?
Ae can lin? any ser"er pro"ided %e !a"e t!e 5L#3.' pro"ider from
(icrosoft to allo% a lin?. @or 5racle %e !a"e a 5L#3.' pro"ider for
oracle t!at microsoft pro"ides to add it as a lin?ed ser"er to t!e sql
ser"er group.
Eo? o you trou<leshoot SQL Ser/er if its runnin# /ery slo??
@irst c!ec? t!e processor and memory usage to see t!at processor is
not abo"e VKZ utilization and memory not abo"e +K3+,Z utilization
t!en c!ec? t!e dis? utilization using 6erformance (onitor, Secondly,
use SQL 6rofiler to c!ec? for t!e users and current SQL acti"ities and
4obs running %!ic! mig!t be a problem. !ird %ould be to run
$6.2#JS2&S&0S command to update t!e inde/es.
What is lo# shippin#?
0an %e do logs!ipping %it! SQL Ser"er *.K 3 Logs!ipping is a ne%
feature of SQL Ser"er 2KKK. Ae s!ould !a"e t%o SQL Ser"er 3
#nterprise #ditions. @rom #nterprise (anager %e can configure t!e
logs!ipping. &n logs!ipping t!e transactional log file from one ser"er is
automatically updated into t!e bac?up database on t!e ot!er ser"er. &f
one ser"er fails, t!e ot!er ser"er %ill !a"e t!e same db and %e can
use t!is as t!e .1 (disaster reco"ery) plan.
Let us say the SQL Ser/er "rashe an you are re<uilin# the
ata<ases in"luin# the master ata<ase ?hat pro"eure to
you follo??
@or restoring t!e master db %e !a"e to stop t!e SQL Ser"er first and
t!en from command line %e can type SQLS#1>#1 .m %!ic! %ill
basically bring it into t!e maintenance mode after %!ic! %e can
restore t!e master db.
What is 2!-? When o ?e use it?
'ul?0opy is a tool used to copy !uge amount of data from tables and
"ie%s. 'ut it %onCt copy t!e structures of t!e same.
What is the ifferen"e <et?een ora"le,sql an sql ser/er B
5racle is based on 1.'(S.
SQL is Structured Query Language.
SQL Ser"er is anot!er tool for 1.'(S pro"ided by (icroSoft.
?hy you nee ine&in# ? ?here that is stroe an ?hat you
mean <y s"hema o<Ne"t? +or ?hat purpose ?e are usin# /ie??
Ae cant create an &nde/ on &nde/.. &nde/ is stoed in userJinde/
table.#"ery ob4ect t!at !as been created on Sc!ema is Sc!ema 5b4ect
li?e able,>ie% etc.&f %e %ant to s!are t!e particular data to "arious
users %e !a"e to use t!e "irtual table for t!e 'ase table...So t!t is a
"ie%.
inde/ing is used for faster searc! or to retrie"e data faster from
"arious table. Sc!ema containing set of tables, basically sc!ema
means logical separation of t!e database. >ie% is crated for faster
retrie"al of data. &t=s customized "irtual table. %e can create a single
"ie% of multiple tables. 5nly t!e dra%bac? is.."ie% needs to be get
refres!ed for retrie"ing updated data.
Differen"e <et?een Store -ro"eure an Tri##er?
%e can call stored procedure e/plicitly.
but trigger is automatically in"o?ed %!en t!e action defined in
trigger is done.
e/- create trigger after &nsert on
t!is trigger in"o?ed after %e insert somet!ing on t!at table.
Stored procedure can=t be inacti"e but trigger can be &nacti"e.
riggers are used to initiate a particular acti"ity after fulfilling
certain condition.&t need to define and can be enable and disable
according to need.
What is the a/anta#e to use tri##er in your -L?
riggers are fired implicitly on t!e tables9"ie%s on %!ic! t!ey are
created. !ere are "arious ad"antages of using a trigger. Some of
t!em are-
Suppose %e need to "alidate a .(L
statement(insert9$pdate9.elete) t!at modifies a table t!en %e
can %rite a trigger on t!e table t!at gets fired implicitly
%!ene"er .(L statement is e/ecuted on t!at table.
2not!er reason of using triggers can be for automatic updation
of one or more tables %!ene"er a .(L9..L statement is
e/ecuted for t!e table on %!ic! t!e trigger is created.
riggers can be used to enforce constraints. @or eg - 2ny
insert9update9 .elete statements s!ould not be allo%ed on a
particular table after office !ours. @or enforcing t!is constraint
riggers s!ould be used.
riggers can be used to publis! information about database
e"ents to subscribers. .atabase e"ent can be a system e"ent
li?e .atabase startup or s!utdo%n or it can be a user e"en li?e
$ser loggin in or user logoff.
What the ifferen"e <et?een ;)39) an ;)39)1LL?
$nion %ill remo"e t!e duplicate ro%s from t!e result set %!ile $nion
all does=nt.
What is the ifferen"e <et?een T0;)!1TE an DELETE
"ommans?
'ot! %ill result in deleting all t!e ro%s in t!e table .1$702# call
cannot be rolled bac? as it is a ..L command and all memory space
for t!at table is released bac? to t!e ser"er. 1$702# is muc!
faster.A!ereas .#L## call is an .(L command and can be rolled
bac?.
Whi"h system ta<le "ontains information on "onstraints on all
the ta<les "reate ?
yes,
$S#1J057S12&7S,
system table contains information on constraints on all t!e tables
created
E&plain normali*ation ?
7ormalisation means refining t!e redundancy and maintain
stablisation. t!ere are four types of normalisation -
first normal forms, second normal forms, t!ird normal forms and
fourt! 7ormal forms.
Eo? to fin out the ata<ase name from SQLJ-L;S "omman
prompt?
Select S from globalJnameP
!is %ill gi"e t!e datbase name %!ic! u r currently connected to.....
What is the ifferen"e <et?een SQL an SQL Ser/er ?
SQLSer/er is an 1.'(S 4ust li?e oracle,.'2 from (icrosoft
%!ereas
Stru"ture Query Lan#ua#e $SQL%, pronounced QsequelQ, is a
language t!at pro"ides an interface to relational database systems. &t
%as de"eloped by &'( in t!e 1)*Ks for use in System 1. SQL is a de
facto standard, as %ell as an &S5 and 27S& standard. SQL is used to
perform "arious operations on 1.'(S.
What is iffren"e <et?een !oBrelate su< query an neste
su< query?
!orrelate su<query runs once for eac! ro% selected by t!e outer
query. &t contains a reference to a "alue from t!e ro% selected by t!e
outer query.
)este su<query runs only once for t!e entire nesting (outer) query.
&t does not contain any reference to t!e outer query ro%.
@or e/ample,
!orrelate Su<query'
select e1.empname, e1.basicsal, e1.deptno from emp e1 %!ere
e1.basicsal \ (select ma/(basicsal) from emp e2 %!ere e2.deptno \
e1.deptno)
)este Su<query'
select empname, basicsal, deptno from emp %!ere (deptno, basicsal)
in (select deptno, ma/(basicsal) from emp group by deptno)
WE1T 9-E01T90 -E0+90MS -1TTE0) M1T!E3)=?
6attern matc!ing operator is L&<# and it !as to used %it! t%o
attributes
1. Z and
2. J ( underscore )
Z means matc!es zero or more c!aracters and under score means
mat!ing e/actly one c!aracter
1%What is ifferen"e <et?een 9ra"le an MS 1""ess?
2% What are isa/anta#es in 9ra"le an MS 1""ess?
3% What are feratures@a/anta#es in 9ra"le an MS 1""ess?
5racle=s features for distributed transactions, materialized "ie%s and
replication are not a"ailable %it! (S 2ccess. !ese features enable
5racle to efficiently store data for multinational companies across t!e
globe. 2lso t!ese features increase scalability of applications based on
5racle.
What is ata<ase?
2 database is a collection of data t!at is organized so t!at itscontents
can easily be accessed, managed and updated. open t!is url -
!ttp-99%%%.%ebopedia.com9#1(9d9database.!tml
What is "luster."luster ine& an non "luster ine& ?
0lustered &nde/-3 2 0lustered inde/ is a special type of inde/ t!at
reorders t!e %ay records in t!e table are p!ysically stored. !erefore
table may !a"e only one clustered inde/.7on30lustered &nde/-3 2
7on30lustered inde/ is a special type of inde/ in %!ic! t!e logical
order of t!e inde/ does not matc! t!e p!ysical stored order of t!e
ro%s in t!e dis?. !e leaf nodes of a non3clustered inde/ does not
consists of t!e data pages. instead t!e leaf node contains inde/ ro%s.
Eo? "an i hie a parti"ular ta<le name of our s"hema?
you can !ide t!e table name by creating synonyms.
e.g) you can create a synonym y for table /
create synonym y for /P
What is ifferen"e <et?een D2MS an 0D2MS?
!e main difference of .'(S M 1.'(S is
1.'(S !a"e 7ormalization. 7ormalization means to refining t!e
redundant and maintain t!e stablization.
t!e .'(S !asn=t normalization concept.
What are the a/anta#es an isa/anta#es of primary :ey
an forei#n :ey in SQL?
-rimary :ey
2d"antages
1) &t is a unique ?ey on %!ic! all t!e ot!er candidate ?eys are
functionally dependent
.isad"antage
1) !ere can be more t!an one ?eys on %!ic! all t!e ot!er attributes
are dependent on.
+orei#n .ey
2d"antage
1)&t allo%s refrencing anot!er table using t!e primary ?ey for t!e
ot!er table
Whi"h ate fun"tion is use to fin the ifferen"e <et?een t?o
ates?
datediff
for #g- select datediff (dd,=23KW32KK*=,=*3KW32KK*=)
output is ,
,K. Go% many clustered inde/es can be created on a tableB
2 clustered inde/ sorts and stores t!e data ro%s in t!e table based on
t!e inde/ ?ey "alues. !erefore only one clustered inde/ can be
created on eac! table because t!e data ro%s t!emsel"es can only be
sorted in one order.
o create a different clustered inde/ on t!e same table, c!ange t!e
!reate as !lustere property setting on t!e e/isting clustered inde/
before creating t!e second inde/.
,1. Go% many non clustered inde/es can be created on a tableB
Fou can create multiple nonclustered inde/es on a table or inde/ed
"ie%. 8enerally, nonclustered inde/es are created to impro"e t!e
performance of frequently used queries not co"ered by t!e clustered
inde/.
7onclustered inde/es are implemented in t!e follo%ing %ays-
61&(21F <#F and $7&Q$# constraints
A!en you create a 61&(21F <#F constraint, a unique clustered
inde/ on t!e column or columns is automatically created if a
clustered inde/ on t!e table does not already e/ist and you do
not specify a unique nonclustered inde/. !e primary ?ey column
cannot allo% 7$LL "alues.
A!en you create a $7&Q$# constraint, a unique nonclustered
inde/ is created to enforce a $7&Q$# constraint by default. Fou
can specify a unique clustered inde/ if a clustered inde/ on t!e
table does not already e/ist. @or more information, see PRIMARY
KEY Constraints and UNIQUE Constraints.
&nde/ independent of a constraint
'y default, a nonclustered inde/ is created if clustered is not
specified. !e ma/imum number of nonclustered inde/es t!at
can be created per table is ))). !is includes any inde/es
created by 61&(21F <#F or $7&Q$# constraints, but does not
include ;(L or spatial inde/es.
7onclustered inde/ on an inde/ed "ie%
2fter a unique clustered inde/ !as been created on a "ie%,
nonclustered inde/es can be created. @or more information, see
Creating Indexed Views.
,2. %!at is stored procedure in sql ser"erB
(icrosoft SQL Serer pro"ides t!e stored !roced"re mec!anism to
simplify t!e database de"elopment process by grouping #ransact$SQL
statements into manageable bloc?s.
2enefits of Store -ro"eures
A!y s!ould you use stored proceduresB Let=s ta?e a loo? at t!e ?ey
benefits of t!is tec!nology-
-re"ompile e&e"ution. SQL Ser"er compiles eac! stored
procedure once and t!en reutilizes t!e e/ecution plan. !is
results in tremendous performance boosts %!en stored
procedures are called repeatedly.
0eu"e "lientCser/er traffi". &f net%or? band%idt! is a
concern in your en"ironment, you=ll be !appy to learn t!at stored
procedures can reduce long SQL %"eries to a single line t!at is
transmitted o"er t!e %ire.
Effi"ient reuse of "oe an pro#rammin# a<stra"tion.
Stored procedures can be used by multiple users and client
programs. &f you utilize t!em in a planned manner, you=ll find t!e
de"elopment cycle ta?es less time.
Enhan"e se"urity "ontrols. Fou can grant users permission
to e/ecute a stored procedure independently of underlying ta&'e
permissions.
Stored procedures are "ery similar to user3defined functions, but
t!ere are subtle differences. @or more information, read Co(!aring
Stored Proced"res and User$)efined *"nctions.
Stru"ture
Stored procedures are e/tremely similar to t!e constructs seen in
ot!er programming languages. !ey accept data in t!e form of input
parameters t!at are specified at e/ecution time. !ese input
parameters (if implemented) are utilized in t!e e/ecution of a series of
statements t!at produce some result. !is result is returned to t!e
calling en"ironment t!roug! t!e use of a recordset, output parameters
and a return code. !at may sound li?e a mout!ful, but you=ll find t!at
stored procedures are actually quite simple.
E&ample
Let=s ta?e a loo? at a practical e/ample. 2ssume %e !a"e t!e table
s!o%n at t!e bottom of t!is page, named &n"entory. !is information
is updated in real3time and %are!ouse managers are constantly
c!ec?ing t!e le"els of products stored at t!eir %are!ouse and
a"ailable for s!ipment. &n t!e past, eac! manager %ould run queries
similar to t!e follo%ing-
S#L#0 6roduct, Quantity
@15( &n"entory
AG#1# Aare!ouse \ =@L=
!is resulted in "ery inefficient performance at t!e SQL Ser"er. #ac!
time a %are!ouse manager e/ecuted t!e query, t!e database ser"er
%as forced to recompile t!e query and e/ecute it from scratc!. &t also
required t!e %are!ouse manager to !a"e ?no%ledge of SQL and
appropriate permissions to access t!e table information.
Ae can simplify t!is process t!roug! t!e use of a stored procedure.
Let=s create a procedure called spJ8et&n"entory t!at retrie"es t!e
in"entory le"els for a gi"en %are!ouse. Gere=s t!e SQL code-
01#2# 6150#.$1# spJ8et&n"entory
alocation "arc!ar(1K)
2S
S#L#0 6roduct, Quantity
@15( &n"entory
AG#1# Aare!ouse \ alocation
5ur @lorida %are!ouse manager can t!en access in"entory le"els by
issuing t!e command
#;#0$# spJ8et&n"entory =@L=
!e 7e% For? %are!ouse manager can use t!e same stored procedure
to access t!at area=s in"entory.
#;#0$# spJ8et&n"entory =7F=
8ranted, t!is is a simple e/ample, but t!e benefits of abstraction can
be seen !ere. !e %are!ouse manager does not need to understand
SQL or t!e inner %or?ings of t!e procedure. @rom a performance
perspecti"e, t!e stored procedure %ill %or? %onders. !e SQL Se"er
creates an e/ecution plan once and t!en reutilizes it by plugging in t!e
appropriate parameters at e/ecution time.
7o% t!at you="e learned t!e benefits of stored procedures, get out
t!ere and use t!em_ ry a fe% e/amples and measure t!e
performance en!ancements ac!ie"ed 33 you=ll be amazed_
&n"entory able
3D -rou"t Warehouse Quantity
1+2 8reen beans 7F 1KK
21+ 6eas @L 2KK
V2, 0orn 7F 1+K
,12 Lima beans 7F 1VK
+)1 omatoes @L VK
3*) Aatermelon @L V,
,1. ypes of Stored 6rocedures
!ere are many types of stored procedures a"ailable in (icrosoft SQL
Ser"er 2KK,. !is topic briefly describes eac! stored procedure type
and includes an e/ample of eac!.
User$defined Stored Proced"res
Stored procedures are modules or routines t!at encapsulate code for
reuse. 2 stored procedure can ta?e input parameters, return tabular
or scalar results and messages to t!e client, in"o?e data definition
language (..L) and data manipulation language (.(L) statements,
and return output parameters. &n SQL Ser"er 2KK,, a stored
procedure can be of t%o types- ransact3SQL or 0L1.
Transa"tBSQL
2 ransact3SQL stored procedure is a sa"ed collection of ransact3SQL
statements t!at can ta?e and return user3supplied parameters. @or
e/ample, a stored procedure mig!t contain t!e statements needed to
insert a ne% ro% into one or more tables based on information
supplied by t!e client application. 5r, t!e stored procedure mig!t
return data from t!e database to t!e client application. @or e/ample,
an e3commerce Aeb application mig!t use a stored procedure to
return information about specific products based on searc! criteria
specified by t!e online user.
!L0
2 0L1 stored procedure is a reference to a (icrosoft .7# @rame%or?
common language runtime (0L1) met!od t!at can ta?e and return
user3supplied parameters. !ey are implemented as public, static
met!ods on a class in a .7# @rame%or? assembly. @or more
information, see 0L1 Stored 6rocedures.
Extended Stored Proced"res
3mportant'
!is feature %ill be remo"ed in a future "ersion of (icrosoft SQL
Ser"er. 2"oid using t!is feature in ne% de"elopment %or?, and plan to
modify applications t!at currently use t!is feature. $se CLR
Integration instead.
#/tended stored procedures let you create your o%n e/ternal routines
in a programming language suc! as 0. #/tended stored procedures
are .LLs t!at an instance of (icrosoft SQL Ser"er can dynamically
load and run. #/tended stored procedures run directly in t!e address
space of an instance of SQL Ser"er and are programmed by using t!e
SQL Ser"er #/tended Stored 6rocedure 26&.
)ote'
0L1 &ntegration pro"ides a more robust and secure alternati"e to
%riting e/tended stored procedures.
S+ste( Stored Proced"res
(any administrati"e acti"ities in SQL Ser"er 2KK, are performed
t!roug! a special ?ind of procedure ?no%n as a system stored
procedure. @or e/ample, sys.sp5"han#e<o?ner is a system stored
procedure. System stored procedures are p!ysically stored in t!e
1esource database and !a"e t!e sp5 prefi/. System stored
procedures logically appear in t!e sys sc!ema of e"ery system3 and
user3defined database. &n SQL Ser"er 2KK,, 8127, .#7F, and
1#>5<# permissions can be applied to system stored procedures. @or
a complete list of system stored procedures, see System Stored
6rocedures (ransact3SQL).
SQL Ser"er supports t!e system stored procedures t!at pro"ide an
interface from SQL Ser"er to e/ternal programs for "arious
maintenance acti"ities. !ese e/tended stored procedures use t!e
&p5 prefi/. @or a complete list of e/tended stored procedures, see
8eneral #/tended Stored 6rocedures (ransact3SQL).
,2. &s it possible to pass a parameter in Stored 6rocedureB
!ere are t%o %ays to pass parameters to a stored procedure using
SQL#/ec. 5ne %ay, %!ic! %or?s across all "ersions of >isual @o/6ro, is
to build t!e SQL command as a string "ariable. !e ad"antage of t!is
met!od is t!at you can c!ec? t!e string and see e/actly %!ic! SQL
command you are passing to t!e bac? end.
!e ot!er %ay is to pass t!e @o/pro "ariables preceded %it! question
mar?s, as in a parameterized "ie%. &n >isual @o/6ro "ersion ,.K and
later "ersions, t!is allo%s you to obtain "alues from t!e stored
procedure t!at are being returned as output parameters.