Stepbysteplearningsasbasicsv1 150218054117 Conversion Gate02

7/21/2019 Stepbysteplearningsasbasicsv1 150218054117 Conversion Gate02

1/121

Step by step

guide to LearnSASVenkat Reddy


2/121

Contents

What is SAS?

Step-1: SAS windowing environment

Step-: SAS !atasets and Variab"es

Step-#: $mporting data into SAS Step-%: &asi' (ro'edures and )un'tions

Step-*: Combining !atasets in SAS

+e,t Steps

Ve

nkata

Reddy

-onasan

i


3/121

&e.ore you start/

(re Re0uisites &asi' data base know"edge

&asi' idea on Ana"yti's and App"i'ations

Windows S with minimum 2& RA3

SAS so.tware

!is'"aimer 4his presentation is 5ust '"ass notes6 4he best way to

treat this is as a high-"eve" summary7

4he a'tua" session went more in depth and'ontained other in.ormation6

3ost o. this materia" was written as in.orma" notes8not intended .or pub"i'ation

Ve

nkata

Reddy

-onasan

i

#


4/121

What is SAS?


5/121

Ana"ysis App"i'ations

(redi'ting what wi"" be the right segment .or amarketing 'ampaign 9 !e'ision 4rees

(redi'ting the "oan repay 'apa'ity o. a 'ustomer 9Logisti' Regression

)ore'asting the revenue numbers .or a produ't -AR$3A

$denti.ying .raud '"aims in hea"th'are insuran'e 9C"uster Ana"ysis

4he re'ommendation engine that predi'ts thebest produ't that interests the user - +eura"networks

(redi'ting the sa"es numbers based on ma'roe'onomi' data 9 Regression ana"ysis

Ve

nkata

Reddy

-onasan

i

*


6/121

What is 2auss markov theorem

Ve

nkata

Reddy

-onasan

i

4he above theorem is used ;nd the regression estimates


7/121

A"" ana"yti's a"gorithms are'omp"i'ated We need pre written "ibraries or 'odes to per.orm

ana"yti'a" operations

&e"ow too"s wi"" make ana"yti's easy SAS

RS(SS

3at"ab


8/121

3ain (hases in "earningana"yti's too"

Ve

nkata

Reddy

-onasan

i

>

1) Basics $ntrodu'tion to the

too" Coding and debugging $mportant .eatures

and "imitations

2) DataHandling $mporting the data !ata manimpu"ations

'reating 'a"'u"ated;e"ds

!ata 3erging

3) Functionsand Algorithms

&asi' and Advan'ed.un'tions

Statisti'a" te'hni0uesand a"gorithms

&asi' Reporting ande,porting the resu"ts


9/121

Step-1: SAS windowingenvironment

Ve

nkata

Reddy

-onasan

i

@


10/121

Contents

$ntrodu'tion to SAS

A''essing SAS


11/121

What is SAS?

Co""e'tion o. modu"es that are used to pro'ess andana"yBe data6

!eve"oped in the early 1970s at North Carolinatate !ni"ersity

rigina""y intended .or management and ana"ysiso. agri'u"tura" ;e"d e,periments

+ow the most wide"y used statisti'a" so.tware

sed to stand .or DStatisti'a" Ana"ysis SystemE8

now it is not an a'ronym .or anything (ronoun'ed DsassE8 not spe""ed out as three

"etters6

Ve

nkata

Reddy

-onasan

i

11


12/121

Various $ndustries useSAS

Casinos

Communi'ations


13/121

A''essing SAS

Ve

nkata

Reddy

-onasani

1#


14/121

A''essing SAS 9 on"inu,Goptiona"H16 4ype sas 6 4his opens the SAS Ddisp"ay managerE8

whi'h 'onsists o. three windows Gprogram8 "og8 andoutputH6 Some pro'edures must be run .rom thedisp"ay manager6

6 4ype sas -nodms 6 Iou wi"" be prompted .or ea'h SASstatement8 and output wi"" s'ro"" by on the s'reen6

#6 4ype sas -stdio 6 SAS wi"" a't "ike a standard +$Jprogram8 e,pe'ting input .rom standard input8sending the "og to standard error8 and the output to

standard output7%6 4ype sas ;"ename6sas 6 4his is the bat'h mode o.

SAS -your program is read .rom ;"ename6sas8 the "oggoes to ;"ename6"og and the output goes to;"ename6"st6

Ve

nkata

Reddy

-onasani

1%


15/121

4he SAS S'reen

Ve

nkata

Reddy

-onasani

1*


16/121

4he SAS S'reen - $

Ve

nkata

Reddy

-onasani

1

Log Window

Explorer

Window

Editor Window

Output Window

Results

Window

(not shown)

Run button click on this button

to run SAS code

Click here or

SAS help!ew Window button

S"#e button

$ool b"r si%il"r

to Windows "pplic"tions


17/121

Lab:

pen SAS .rom start button6 $denti.y be"ow windows in SAS


18/121


19/121


20/121

Log window

Ve

nkata

Reddy

-onasani

Where is "og window in SAS?

16 4he !ebug window6 !isp"ays the messages

about SAS session and

the program that issubmitted6

Write this s'ript in editor andsee the "og ;"e7

procprint

data=sashelp.airr;

run;


21/121

utputs window

Ve

nkata

Reddy

-onasani

1

16 !isp"ays the "isting output in theoutput window

6 $t automati'a""y opens or moves tothe .ront o. disp"ay when output is'reated6

Write this s'ript in editor and seeoutput ;"e7

procprintdata=sashelp.air;run;


22/121

Resu"ts window

Ve

nkata

Reddy

-onasani

16 Fe"ps navigate and manageoutput

6 View8 save8 and print individua"items o. output6


23/121

Comments and Fe"p menu

Ve

nkata

Reddy

-onasani

#

4here are two sty"es o. 'omments you 'an use:16 ne starts with an asterisk GKH and ends with a

semi'o"on G7H6

6 4he other sty"e starts with a s"ash asterisk GKH andends with an asterisk s"ash GKH6

#ther o$tions% Resu"ts in htm"

output


24/121

Lab: 3y ;rst SAS (rogram

3y ;rst SAS program9 DFe""o wor"dE

pro' print dataMsashe"p6air7

run7

Submit the program

Write this SAS s'ript in your editordataincome_data;

Inputincome expenses;

Cards;

1200 1000

9000 600

;run;

Procprintdata=income_data;

Run;

Ve

nkata

Reddy

-onasani

%


25/121

Lab


26/121

Lab: 3y ;rst SAS program

Run be"ow 'ode

pro' print dataMsashe"p6airr7

run7

Fow to diagnose 'orre't errors

Run be"ow 'odepro' 'ontent dataMsashe"p6air7

run7

$denti.y the errors i. any

(rint in'omeNdata C"ose SAS open it again

(rint in'omeNdata

Ve

nkata

Reddy

-onasani


27/121

SAS Libraries

&y de;ning a "ibrary8 you indi'ate the "o'ation o. yourSAS ;"es to SAS

Co""e'tion o. SAS ;"es su'h as SAS data sets and'ata"ogs

4o a''ess a "ibrary8 you need to assign it a name Ga"so

known as a "ibre.8 or "ibrary re.eren'eH A.ter assigning a "ibrary name youO"" work with SAS

data sets in a "ibrary $n the Windows and ni, environments8 SAS "ibrary is

typi'a""y a group o. SAS ;"es in the same .o"der or

dire'tory6 Fow to de;ne a "ibrary? sing $ntera'tive window Libname Statement

Ve

nkata

Reddy

-onasa

ni

=


28/121

4emporary and permanent"ibraries

Ve

nkata

Reddy

-onasa

ni

>

&em$ 'i(rary Storing ;"es temporari"y

$. you donOt spe'i.y a "ibraryname when you 'reate a

;"eor

$. you spe'i.y the "ibraryname WR

4hen the ;"e is stored in the

temporary SAS data "ibrary

Last on"y .or the 'urrentSAS session

ermanent li(rary Storing ;"es

permanent"y

Spe'i.y a "ibrary nameother than the de.au"t"ibrary name Work

Avai"ab"e duringsubse0uent SASsessions


29/121

!e;ning a "ibrary -2$

!e;ne a "ibrary name usingintera'tive window n the too"bar8 '"i'k the +ew Library too"


30/121

Libname Statement

Write a statement in a editor window

Synta,

L$&+A3< "ibre. OSAS-"ibraryQ 7


31/121

Ru"es .or assigning aLibrary16 Limited to > 'hara'ters "ong

6 3ust start with "etter orunders'ore on"y

#6 Can be a 'ombination o. "etters8numbers and unders'ore

Venkata

Reddy

-onasa

ni

#1


32/121

Re.eren'ing ;"es in SASLibrariesAs we know8 to re.eren'e a SAS ;"e we need to assign a libref

G"ibrary re.eren'eH to the SAS "ibrary in whi'h the ;"e is

stored6 Re.eren'ing ;"es .rom permanent "ibrary:

4o re.eren'e permanent SAS data set8 you use two-"eve"

name6 libref.lename 4he libref is name o. the SAS data "ibrary that 'ontains the

;"e6 4he lenameis the name o. the ;"e itse".6

A period separates the librefand lename6


33/121

SAS de.au"t "ibraries GSe".H

)o""owing "ibraries are automati'a""y assigned ea'h timeyou start SAS

Sashe"p: (ermanent "ibrary Contains samp"e data and other ;"es that 'ontro"

how SAS works at our site6 4his is a read-on"y "ibrary

Sasuser: (ermanent "ibrary Contains SAS ;"es that store persona" settings our

own ;"es

Work: 4emporary "ibrary Last on"y .or the 'urrent SAS session

Venkata

Reddy

-onasa

ni

##


34/121

Lab

(rint a data ;"e .rom SAS de.au"t "ibrary

(ro' print dataM UP 7 run7

(rint a data ;"e .rom SASuser "ibrary

(rint a dataset .rom work "ibrary6

4ry work6 without work6 Create a new "ibrary DNdataNE

Create a new "ibrary D#dataNE

Create a new "ibrary DdataE

Venkata

Reddy

-onasa

ni

#%


35/121

Step-: SAS datasetsand variab"es

Venkata

Reddy

-onasa

ni

#*


36/121

Contents

SAS programs

!ata step

(RC Step

Writing a SAS program and debugging it

SAS !ata sets !ata sets properties

Variab"es in a data set

4ypes o. variab"es

Attributes o. variab"es

Create data using !ata statement Venkata

Reddy

-onasa

ni

#


37/121

SAS (rogram

Se0uen'e o. statements e,e'uted in order

sed to a''ess8 manage8 ana"yBe and present thedata

Contains statements8 e,pressions8 .un'tions and

CALL routines8 options8 .ormats8 and in.ormats Simp"i;ed programming with bui"t in programs

known as A rocedures Gpre written 'odesH (RC R


38/121

Chara'teristi's and "ayout

SAS programs are made up o. SAS statements6

A SAS statement has two important'hara'teristi's:

16 $t usua""y begins with a SAS keyword66 $t a"ways ends with a semi'o"on

Layout o. SAS program:16 4hey 'an begin and end anywhere on a "ine

6 ne statement 'an 'ontinue over severa" "ines

#6 Severa" statements 'an be on a "ine6

%6 SAS statements are +4 'ase sensitive6

*6 &"anks or spe'ia" 'hara'ters separate thewords in a SAS statement6

Venkata

Reddy

-onasa

ni

#>


39/121

Lab

(rint a dataset in SAS he"p "ibrary

(ro' print dataMUP7 run7

4ry these properties o. SAS statements

16 4hey 'an begin and end anywhere on a "ine -

4ry it6 ne statement 'an 'ontinue over severa"

"ines - 4ry it

#6 Severa" statements 'an be on a "ine6 - 4ry it

%6 SAS statements are +4 'ase sensitive6 - 4ry

it Venkata

Reddy

-onasa

ni

#@

C t . SAS


40/121

Components o. SAS(rograms

SAS programs are 'onstru'ted .rom two basi' bui"dingb"o'ks:

16 !A4A steps

6 (RC steps 4hese two types o. steps8 a"one or 'ombined8 .orm a""

SAS programs6 2enera""y8 a step ends with a R+ statement or when a

new !A4A or (RC step begins6

Venkata

Reddy

-onasa

ni

%


41/121

!A4A step

&egins with the keyword !A4A

Creates or modi;es the data set

(rodu'es 'ustom designed reports

sing data step we 'an

(ut our data into a SAS data set

Compute va"ues 'reate new ;e"ds

Che'k .or and 'orre't errors in our data

(rodu'e new SAS data sets by sub setting8 merging8and updating e,isting data sets

Venkata

Reddy

-onasa

ni

%1


42/121

Creating a SAS data

data U"ibnameP6UdatasetnameP7

$nput Uvar1P UvarP/6UvarpP7

Cards7

v11 v1 /6vp1

////66v1n vn /6vpn

7

run7

Venkata

Reddy

-onasa

ni

%


43/121

Creating a SAS data

V

enkata

Reddy

-onasa

ni

%#

datamylib.stock;

inputolume price;

cards;

1200 1!

1!00 1!."

12"0 161!#0 1$."

1290 1%

1609 19."

1%09 1#

112$ 16.";

run;

! t ti . SAS d t


44/121

!ata portion o. SAS dataset

bservations GRowsH: Co""e'tions o. data va"ues

that usua""y re"ate to asing"e ob5e't

Can store any number o.

observations6

Variab"es GCo"umnsH: Co""e'tions o. va"ues that

des'ribe a parti'u"ar

'hara'teristi' Can store thousands o.

variab"es6

V

enkata

Reddy

-onasa

ni

%%


45/121

Lab

se data step and 'reate a dataset namedDrea""ybigdatasetnamethis1#%*=>E7

se data step and 'reate a dataset namedD1#dataE7

se data step and 'reate a dataset namedDTdataE7

se data step and 'reate a dataset namedDNdataE7

se data step and 'reate a dataset namedDNdata1#E7

V

enkata

Reddy

-onasa

ni

%*


46/121

SAS !ataset name ru"es

16 Can be 1 to # 'hara'ters "ong

6 3ust begin with a "etter GA98 either upper'ase or"ower'aseH or an unders'ore G N H

#6 Can be a 'ombination o. numbers8 "etters8 orunders'ores6


47/121

Lab

se data step and 'reate a variab"e namedDrea""ybigvarnamethis1#%*=>@1111E7 in adataset

se data step and 'reate a variab"e named

D1#varE7 se data step and 'reate a variab"e named

DTvarE7

se data step and 'reate a variab"e namedDNvarE7

se data step and 'reate a variab"e namedDNvar1#E7

V

enkata

Reddy

-onasa

ni

%=


48/121

SAS Variab"es

Co"umns in the SAS datasets are 'a""ed SASvariab"es6Attributes o. SAS variab"es: Attribute in.ormation in'"udes

the variab"eOs

16 4ype

6 Length

#6 )ormat

%6 $n.ormat

*6 Labe"6

SAS variab"e name ru"es: GSame as SAS dataset nameH

6 Limited to # 'hara'ters "ong=6 3ust start with "etter or unders'ore

>6 Can be a 'ombination o. "etters Gpper'ase "ower'ase8A-H 8 numbers and unders'ore

V

enkata

Reddy

-onasa

ni

%>

L b C ti "


49/121

Lab: Creating a samp"edataset using s'ript Run be"ow s'riptDatastudent;

inputhei&ht 'ei&ht;

Cards;

120 ""

1$$ #01!0 #%

1$" 69

;

Run;

See the properties o. the 'reated data8 pro' 'ontents

Add a new variab"e Qname to the above dataset

What is the error? Gse TH

See pro' 'ontents data

(rint data

V

enkata

Reddy

-onasa

ni

%@

3ain Attributes o. a


50/121

3ain Attributes o. avariab"e Length:

Chara'ter variab"es 'an be up to #8== bytes "ong6 A"" numeri' variab"es have a de.au"t "ength o. >6

Labe": Consists o. des'riptive te,t up to * 'hara'ters "ong Labe"s are used instead o. variab"e names in some reports

Labe"s are used .or the 'o"umn headings in the V$


51/121

3ain types o. SASvariab"es

Chara'ter variab"es:16 Can 'ontain any va"ue

6 se &"ank to represent the missing va"ues

#6 Can be up to #

+umeri' variab"es:%6 Can 'ontain on"y numeri' va"ues

G digits through @8 Y8-868 < or s'ienti;' notationsH

6 se a sing"e period G6H to represent the missing va"ues

#6 Fave a de.au"t "ength o. >6

%6 +umeri' va"ues are stored as Zoating point number as> bytes o. storage8 un"ess you spe'i.y another "ength6 V

enkata

Reddy

-onasa

ni

*1


52/121

Step-#: $mporting datainto SAS

V

enkata

Reddy

-onasa

ni

*


53/121

Contents

!ata set 'reation using SAS programGdata stepH

sing import wiBard $mporting


54/121

Reading data using importwiBard

V

enkata

Reddy

-onasa

ni

*%

&

'


55/121

Lab

$mport SurveyNdata .rom SurveyNdata6,"s ;"e

+ame the output ;"e as surveyN17

$mport the ;"e into work "ibrary 'urrent "ibrary

(rint the 'ontents o. the data

Are a"" the variab"es present in the SAS dataset? $mport QCustNsurveyNo"d data .rom

SurveyNdata16,"s

V

enkata

Reddy

-onasa

ni

**


56/121

$mport CSV and 4A& ;"es

V

enkata

Reddy

-onasa

ni

*

&

'


57/121

Lab

$mport '"ientNdata6'sv ;"e

+ame the output ;"e as '"ientN17

$mport the ;"e into work "ibrary 'urrent "ibrary


Are a"" the variab"es present in the SAS dataset? $mport '"inetNmanager6t,t ;"e

$mport pri'e web6'sv ;"e

See the 'ontents

a"e the im$ort code in local *older Venkata

Reddy

-onasa

ni

*=

R di d t t i Q(RC


58/121

Reading data sets using Q(RCimport

ROC *+OR$ ,A$A-*LE. /0location of file12s"%ple3d"t"4cs#5

O6$. testd"t4excelSAS&

,7+S.CS8

RELACE9

R6!9

Where

DATAFILEis the n"%e "nd the loc"tion o ile :ou w"nt to re"d4

OUTis the n"%e o the SAS d"t" set :ou w"nt to cre"te

DBMS is identifier; speciies the t:pe o d"t" to i%port4 8"lid identiiers ordeli%ited iles "re ,L+;

Lab: sing (ro' $mport


59/121

Lab: sing (ro' $mports'ript $mport Fea"th'are data


Are a"" the variab"es present in the SAS dataset?

4ry with and without rep"a'e options

$mport 'ars data $mport 'redit risk6t,t data

V

enkata

Reddy

-onasani

*@


60/121

Working with !& ;"es

V

enkata

Reddy

-onasani

What i. we have a"" these tab"es in one database?

1


61/121

Lab

Create a "ibrary out o. A""NSurveyN!&6mdb

(rint the 'ontents o. the ;"es

Copy re"evant tab"es into your "ibrary

(rint '"ientNmanager data

4ry opening a''ess ;"e using a''ess now

V

enkata

Reddy

-onasani

1

Reading data .rom SAS


62/121

Reading data .rom SASdatasets =ow to cre"te " new SAS d"t" set out o SAS d"t" set>

=ow do we cre"te sur#e:' ro% sur#e: d"t">

,"t" libn"%e4new9

Set libn"%e4old9

Run9

V

enkata

Reddy

-onasani

Lab: Creating new


63/121

Lab: Creating newvariab"es Create a new dataset and add a new variab"e

s'ore1 GResponse Y [ua"ityHdata'ork.surey_2;

set'ork.surey;

score1=(response)*uality+,2;

run;

Create a new variab"e diX8 the diXeren'e betweenovera"" rating and s'ore1

pdate the same dataset with a new variab"einterNs'ore whi'h is average o. Communi'ation8Response8 [ua"ity s'ore

(rint a"" the observations with overa"" s'ore graterthan than or e0ua" to %

V

enkata

Reddy

-onasani

#


64/121

3anipu"ating SAS datasets Variab"es

V

enkata

Reddy

-onasani

%


65/121

Contents

Re'ap

Creating and managing variab"es

!rop keep statements

Conditiona" data pro'essing

SAS .un'tions

V

enkata

Reddy

-onasani

*

Lab: Creating new


66/121

Lab: Creating newvariab"es Create a new dataset and add a new variab"e

s'ore1 GResponse Y [ua"ityHdata'ork.surey_2;

set'ork.surey;

score1=(response)*uality+,2;

run;

Create a new variab"e diX8 the diXeren'e betweenovera"" rating and s'ore1

pdate the same dataset with a new variab"einterNs'ore whi'h is average o. Communi'ation8Response8 [ua"ity s'ore

(rint a"" the observations with overa"" s'ore graterthan than or e0ua" to %

V

enkata

Reddy

-onasani


67/121

Re'ap

Create a new "ibrary

What are the a"ready e,isting datasets in your"ibrary?

$mport mar+et,data,onedata into your "ibrary

using import s'ript (rint the 'ontents o. the data

Fow many observations are there in the dataset?

Fow many variab"es?

$denti.y non numeri' variab"es

(rint ;rst ten observations o. the data

Create a new ;e"d genera"N0uestions6 $. wesubtra't numN'ustomN0uestions .romnumN.ormN;e"ds8 we get genera"N0uestions

V

enkata

Reddy

-onasani

=

Creating new variab"e -


68/121

Creating new variab"e -using i. then e"seCreating a new variab"e using i. then e"se

datadatalib.market_one_1;

setdatalib.market_one;

i-bud&et = 3000thenbud&et_ind=/lo'/;

elsei-bud&et = 100000thenbud&et_ind=/hi&h/;elsebud&et_ind=/medium/;

run;

V

enkata

Reddy

-onasani

>

Lab: Creating a variab"e-


69/121

Lab: Creating a variab"e-using i. then e"se Create a new variab"e assetNind6

$. the number o. assets are greater than Bero thenassetNind takes the va"ue yes otherwise +

See pro' 'ontents

(rint ;rst ten observations Create rea'hNind whi'h takes va"ues 18 and #

when rea'h is "ess than ##8 ## to and = to16

(rint ;rst ten observationsV

enkata

Reddy

-onasani

@


70/121

!rop and keep variab"es

What i. $m not interested in a"" the ;e"ds? Creating a new data set with "ess number o.

;e"ds6

sing drop statementdatadatalib.market_one_2;

set datalib.market_one (drop=name start_date end_date+ ;

run;

sing eep statementdatadatalib.market_one_$;

set datalib.market_one (keep=id name bud&et+ ;

run;

V

enkata

Reddy

-onasani

=

Lab: !rop and keep


71/121

Lab: !rop and keepvariab"es $mport marketNdk data See the 'ontents

Var18 var8 var# are o. no use .or the ana"ysis

Create a new data by dropping the unne'essary

variab"es by using drop statement Create a new data by keeping on"y ne'essary

variab"es by using keep statement

V

enkata

Reddy

-onasani

=1


72/121

Sub setting the data

Sometimes we need a subset o. the data .or.urther ana"ysis

datadatalib.market_one_!;

setdatalib.market_one;'hereertical=/echnolo&y/;

run;

or

datadatalib.market_one_!;


i-ertical=/echnolo&y/;

run;

V

enkata

Reddy

-onasani

=


73/121

Lab Sub-setting the data

Create a data set named DgoodNrea'hN'ampE'ontains a"" the 'amps where rea'h is more than *per'ent

Create using where statement

Create the same using i. statement $s there any diXeren'e between where and i.?

(rint a subset o. data: 4ry to print the data where rea'h is more than @

se i. and print again

i. +4%# observations read .rom the data set !A4AL$&63AR


74/121

3a5or diXeren'esbetween where and i.

V

enkata

Reddy

-onas

ani

=%

Where statement IF statement

Selects observations before they arebrought into the program data vector,making it a more efficientprogramming technique.

Works on observations after they areread into the program data vector.

Cannot be executed conditionally aspart of an I statement

!he sub setting I statement can

be executed conditionally

Can be used in S"S procedures Cannot be used in S"S procedures tosubset observations for bro#sing orediting.


75/121

SAS .un'tions - +umeri'

Sum8 min8 ma,8 avg .un'tions

data334I5.378_t'o_1;

set334I5.378_t'o;

sum_t'o=:um(ebinar


76/121

Lab: )un'tions

16 4here is some error in the number o. assets;e"d6

6 Create a variab"e numNasset to veri.y number o.assets .ormu"a6 +umber o. assents is nothing

but the sum o.WhiteN(aper8Webinar8So.twareN!own"oad8)reeNXer8LiveN


77/121

String .un'tions

datadatalib.market_one_";


ertical_ind=substr(ertical


78/121

Lab: String .un'tions

Create nameNnew by taking ;rst 'hara'ters o.the name variab"e

Create a new variab"e by 'onverting the name o.

the 'ampaign into upper'ase Guse +C"S-

function) Create a Zag variab"e that takes va"ue 1 $.

'ampaign starting month is not e0ua" to 'ampaignending month

Venkata

Reddy

-onas

ani

=>

! ) i


79/121

!ate )un'tions

datadatalib.market_one_";


start_month=month(start_date+;

start_date1=day(start_date+;

uration_days=I>C8(/day/


80/121

Lab: !ate .un'tions

Fow many 'ampaigns started in \anuary? Fow many 'ampaigns started on Saturday or

Sunday?

What is the duration o. ea'h 'ampaign? Create

three variab"es !urationNdays

!urationNmonths

!urationNweeks

Fow many market 'ampaigns run more than

monthsVenkata

Reddy

-onas

ani

>


81/121

Step-%: &asi' (ro'eduresand )un'tions

Venkata

Reddy

-onas

ani

>1

C t t


82/121

Contents

4he (RC step (RC C+4


83/121

Components o. SAS(rograms

SAS programs are 'onstru'ted .rom two basi' bui"dingb"o'ks:

16 !A4A steps

6 (RC steps 4hese two types o. steps8 a"one or 'ombined8 .orm a""

SAS programs6 2enera""y8 a step ends with a R+ statement or when a

new !A4A or (RC step begins6

Venkata

Reddy

-onas

ani

>#

(RC t


84/121

(RC step

&egins with the keyword (RC (re-written routines that enab"e you to ana"yBe

and pro'ess the data

Sometimes 'reate SAS data sets that 'ontain the

resu"ts o. the pro'edure (resent the data in the .orm o. a report

)or e,amp"e Create a report that "ists the data

(rodu'e des'riptive statisti's Create a summary report

(rodu'e p"ots and 'harts

Run regression ana"ysis

Venkata

Reddy

-onas

ani

>%

( C t t


85/121

(ro' Contents

(ro' 'ontents des'ribes the stru'ture o. the SAS data set6 2ives.o""owing in.ormation -

!ata set "eve" !etai"s o. data "ike +ame8 'reation date8 +umber o. observations

G weighted and un-weightedH8 +umber o. variab"es8 )i"e siBe anda''ess permissions

Variab"e "eve" (rovides des'riptive stats a'ross ea'h variab"e su'h as +ame8 4ype

GChara'ter vs6 +umeri'H8Length8 )ormats8 (osition and Labe"s

(ro' 'ontents dataM UUdata namePP7 run7

Useful options :

Short Outputs the list of variables in a row by row format.

Code : proc contents data=test short;run;

Out=filename- Creates a data set wherein each observation is a variable

from the original data set.

Code : proc contents data=test out=test2noprint;run;

Venkata

Reddy

-onas

ani

>*

L b ( C t t


86/121

Lab: (ro' Contents

$mport (ri'eNwebNdata6'sv7 See the datades'ription6pd. ;"e

(rint the 'ontents

Are a"" the variab"es in e,pe'ted .ormat?

(rint on"y ;e"d names o. the datasetproccontentsdata=class1.market_one short;

run;

4ake the output o. 'ontents into a sas ;"e8 name it'ontentsNoutproccontentsdata=class1.market_one out=con_out noprint;

run;Venkata

Reddy

-onas

ani

>

( S t


87/121

(ro' Sort

(rimari"y used to sort the observation o. yourdata by a 'ertain variab"e or 'o""e'tion o.variab"es6

Fowever8 it 'an a"so be used to 'reate a new

data set8 subset your data8 rename8 drop8 orkeep variab"es8 and .ormat or "abe" variab"es6

procsortdata=class1.market_one out=ne';

by start_date ;

run;

A"most a"ways a good idea to use the 4M option when using pro'sort to do anything e,'ept .or a simp"e sort6 &e'ause (ro' Sortautomati'a""y writes over your data set]

Venkata

Reddy

-onas

ani

>=

L b ( S t


88/121

Lab: (ro' Sort

Sort marketNone data by budget Sort marketNone data by budget des'ending

Sort data by date and rea'h

se marketNone tab"e and 'reate a new tab"e

'ampanNnames8 sort name in a"phabeti'a"

Venkata

Reddy

-onas

ani

>>

( S t


89/121

(ro' Sort

Sorting and removing the dup"i'ates

procsortdata=class1.market_one out=class1.names_table nodupkey;

byname ;

run;

Sorting and removing the dup"i'ate re'ords

procsortdata=class1.market_one out=class1.names_table nodup;

byname ;

run;Venkata

Reddy

-onas

ani

>@

Lab (ro' Sort


90/121

Lab: (ro' Sort

Sort the dataset by date7 dont overwrite theorigina" data

Create a new dataset whi'h 'ontains a"" theuni0ue brands on a given dateGA dataset wherethe brand name does not repeat on a given dayH

Fow many uni0ue produ'ts are there in thedataset? Crate a new data set with this uni0ue "ist

Venkata

Reddy

-onas

ani

@

(ro' S0"


91/121

(ro' S0"

procsql;

createtableclass1.market_s*l as

select?

-romclass1.arket_t'o

'herenum_assets0;

quit;

[uit instead o. run?Venkata

Reddy

-onas

ani

@1

Lab (ro' S0"


92/121

Lab: (ro' S0"

$mport marketNdataNthree Crate a new tab"e where budget is greater than Bero

$n marketNone data ;nd the average budgetandaverage rea'h .or ea'h verti'a"

Create a new tab"e by inner 5oining marketNone8marketNtwo 5oin them on idprocsql;

createtableclass1.market_one_t'o as

select?

-romclass1.arket_one asa [email protected]_t'o asb

ona.id=b.id;

quit;

Create a new tab"e by inner 5oining above tab"e tomarketNthree tab"e8 save the ;na" tab"e asmarketN;na"Ndata

Venkata

Reddy

-onas

ani

@

Lab: (ro' S0"


93/121

Lab: (ro' S0"

)ind the .re0uen'y o. ea'h brandG'ount number o.times a brand appearsH8a"so average "ist pri'e .orea'h brand6 Save this data as brandNdata

procsql;

createtablebrand_data as

selectbrand< count(ro'id+ as-re*< mean(listprice+ asa&_list_price

-romclass1.price_data

&roupby brand;

quit;

)or ea'h produ't ;nd the number o. appearan'es8

average rating and tota" review 'ount8 name thisdata as produ'tNdata

Venkata

Reddy

-onas

ani

@#

(ro' 2p"ot


94/121

(ro' 2p"ot

S'atter p"ot

procgplotdata= class1.market_-inal_data;

plotbud&et_leads?bud&et;

'herebud&et 100000;

run;

Venkata

Reddy

-onas

ani

@%

Lab: (ro' 2p"ot


95/121

Lab: (ro' 2p"ot

$. Average pri'e is "ess8 then .re0uen'y wi"" bemore8 i. pri'e is more then .re0uen'y wi"" be "ess6

Veri.y the above statement by drawing a s'atterp"ot graph between .re0 and average "ist pri'e inbrandNdata

Venkata

Reddy

-onas

ani

@*

(ro' 2'hart


96/121

(ro' 2'hart

Verti'a" &ar 'hartprocgchartdata= class1.market_-inal_data;

barertical , type=sumsumar=total_leads;

Run;

Vbar J a,is 'atogiry

4ype type o. aggregation

sumvar I a,is variab"e 4he above s'ript shows sum o. "eads .or ea'h verti'a"

procgchartdata= class1.market_-inal_data;

barertical , type=meansumar=total_leads;

Run;

Venkata

Reddy

-onas

ani

@

Lab: (ro' 2'hart


97/121

Lab: (ro' 2'hart

!raw a verti'a" bar graph whi'h shows a"" thebrands and the average number o. reviewsGseorigina" pri'eNdataH

!raw a horiBonta" bar graph .or the same

!raw a bar 'hart .or number o. produ'ts by sitename Iou may want to use 'ount

4ry .re0 i. it doesnt work

Venkata

Reddy

-onasani

@=

(ie 'hart


98/121

(ie 'hart

PROCGCHART33=class1.price_data;

Icate&ory;

RUN;

Venkata

Reddy

-onasani

@>

Lab: (ie 'hart


99/121

Lab: (ie 'hart

4here was a 'omp"aint that one in every .ouritems "isted on the site is out o. sto'k6 $s thattrue? !raw a pie 'hart to veri.y

!raw a pie whi'h shows the per'entage o.produ'ts with .ree shipping avai"abi"ity nonavai"abi"ity

Venkata

Reddy

-onasani

@@


100/121

Step-*: Combining

!atasets in SAS

Venkata

Reddy

-onasani

1

Co%bing d"t" set


101/121

Co%bing d"t" set

Co%bining d"t"sets in " ,A$A step

oConc"ten"te Append

SE$

o*nterle"#eSE$ ? 7@

o+erge

+ERE

+ERE ? 7@ .enkata/eddy0onasani

(1(

Conc"ten"tionBAppending


102/121

Conc"ten"tionBAppending

Si%pl: "ppendingBst"king o the d"t"

Eg +onth b: %onth d"t" conc"ten"tedto or% :e"r end d"t"

Conc"ten"ting d"t" ro% #"riouscusto%ers to %"ke " %"ster d"t"set

.enkata/eddy0onasani

(1&

!emo: Appending


103/121

!atasetsdatastore1;

inputcustomer A;

cards;

rB

rB3rBD

;

datastore2;

inputcustomer A;cards;

rB5

rBE

rBC

dataoerall;

setstore1 store2;run;

procprintdata=oerall;

run;

.enka

ta/eddy0onasani

(12

Lab: Con'atenation


104/121

Lab: Con'atenation

Create there datasets out o. pri'e web data8 based on brandSamsung8 app"e others

Crate a new variab"e in app"e dataset8 dis'Npri'eNapp"e whi'his ^ o. a'tua" 'ost

Con'atenate the three datasets

Fow many ;e"ds are there in the new dataset? Crate .our datasets out o. te"e'om 'omp"aints data8 based onstatus '"osed8 open8 4emporary C"osed and others

eep on"y de" number8 'ustomer emai" id and status

Con'atenate a"" these .our datasets to .orm a newNovera""dataset

datatech_mark ener_mark;


i-ertical=/echnolo&y/thenoutputtech_mark ;

elsei- ertical=/ner&y/thenoutputener_mark;

run;

.enka

ta/eddy0onasani

(13

*nterle"#e


105/121

*nterle"#e

Appending si%pl: "ppends We %": w"nt to %"int"in the order r"ther th"n Dust "ppending

*nterle"#e Ordered st"cking

.enka

ta/eddy0onasani

(14

!emo: $nter"eave


106/121

!emo: $nter"eave

datastore1;inputcustomer A;

cards;

rB

rB3

rBD;

datastore2;

inputcustomer A;

cards;rB5

rBE

rBC

;

dataoerall;

setstore1 store2;

by name;run;

procprintdata=oerall;run;

.enka

ta/eddy0onasani

(15

!emo: $nter"eave


107/121

!emo: $nter"eave

datastore1;inputcustomer A;

cards;

rB

rB3

rBD;

datastore2;

inputcustomer A;

cards;rB5

rBE

rBC

;

dataoerall;

setstore1 store2;

run;

procprintdata=oerall;

run;

procsortdata=store1;

by customer;

run;

procsortdata=store2;

by customer;

run;

.enka

ta/eddy0onasani

(16

LA& $nter "eaving !ataset


108/121

LA& $nter "eaving !ataset

Crate .our datasets out o. te"e'om 'omp"aints data8based on status '"osed8 open8 4emporary C"osed andothers

eep on"y de" number8 'ustomer emai" id and status6

$nter"eave by a"" these .our datasets by de" number to

.orm a newNovera"" dataset Create there datasets out o. pri'e web data8 based on

brand Samsung8 app"e others

Crate a new variab"e in app"e dataset8 dis'Npri'eNapp"ewhi'h is ^ o. a'tua" 'ost

Con'atenate the three datasets based on date

Fow many ;e"ds are there in the new dataset?.enka

ta/eddy0onasani

(17

)rom set to 3erge


109/121

)rom set to 3erge

3erging is simi"ar to 5oins but not e,a't"y same as 5oins ptions need to be used in proper order to .orm the desired

datasetdatastudents1;

inputname A maths;cards;

7am #%

7obert 90

7aheem %0

Fopi #"3nil 60

;

datastudents2;

inputname A en&lish;

cards;

7am ""7obert #0

7aheem 60

Gred #"

3lex "0

;proc

sort

data=students1; byname;run

;procsortdata=students2; byname; run;

datat'o_sub;

setstudents1 students2;byname;

run;

roc print data=t'o_sub; run;

.e$lace et (y /erge

.enka

ta/eddy0onasani

(18

+"tched +erging


110/121

+"tched +erging

datat'o_sub;

setstudents1 students2;byname;

run;

datat'o_sub;

er&e students1 students2;

byname;run;

Obs name maths english

1 Alex . 50

2 Anil 60 .

3 Fred . 75

4 Gopi 75 .

5 Raheem 80 60

6 Ram 78 55

7 Robert 0 70

Obs name maths english

1 Alex . 50

2 Anil 60 .

3 Fred . 75

4 Gopi 75 .

5 Raheem 80 .

6 Raheem . 60

7 Ram 78 .

8 Ram . 55

Robert 0 .

10 Robert . 70

.enka

ta/eddy0onasani

((1

3erging


111/121

3erging

3erge "ooks .or the mat'hing variab"es and 'reated a'onso"idated merged dataset

!A4A output-SAS-data-set7 3


112/121

LA&: 3at'hed 3erging

$mport rders data spot data into two diXerent datasets .rom 4V 'ommer'ia" data

Sort both data sets based on iSC$A!-i!

3erge orders and s"ots based on iSC$A!-i!

Fow many observations are there in the 'onso"idateddataset

datat'o_sub;

er&e students1 students2;

byname;

run;

.enka

ta/eddy0onasani

((&

3erge With Conditions


113/121

3erge With Conditions

What i. we want 'omp"ete detai"s .rom dataset-1 and mat'hing detai"s .ormdataset-1

3arks o. on"y students .rom data set1 and .et'h their


114/121

'onditiondatat'osub_student1;

mer&estudents1(in=a+ students2(in=b+;

byname;

i-a;

run;

datat'osub_student2;


byname;

i-b;

run;

procprintdata= t'osub_student1;run;

procprintdata= t'osub_student2; run;

.enka

ta/eddy0onasani

((3

3erge Condition


115/121

Statementsust "e in 1st dataset#

i$in1# % &a'e as( i$in1 ) 1#

ust "e in 2nd dataset#

i$in2#

ust "e in "oth datasets#

i$in1 and in2#

.enka

ta/eddy0onasani

((4

3erge Condition


116/121

Statements

in( $ (

in& $ 1

9"!"S-! ( 9"!"S-! &

in( $ 1

in& $ (

in( $ (

in& $ (

.enka

ta/eddy0onasani

((5

LA&: 3erge withdi i


117/121

Condition 3erge the data sets orders and spot data Create a new data set with a"" orders data a"ong

with mat'hing spot data ;e"ds

Create a new data set with a"" spot data a"ongwith mat'hing orders data ;e"ds

((6

.enka

ta/eddy0onasani

datat'osub_student1;


byname;i-a;

run;b

LA&: 3erge withC di i


118/121

Condition Sort both bi"" and 'omp"aints data on de"Nno6

Remove dup"i'ates using nodupkey on de"Nno

Atta'h 'omp"aint detai"s o. the 'ustomers to thebi""ing data8 a"" 'ustomers might not have

'omp"aints Atta'h bi""ing detai"s to 'omp"aints data8 i.avai"ab"e

Create a data set and print a"" the 'ustomerswhose bi""ing and 'omp"aints detai"s are avai"ab"edatat'osub_student1;


byname;

i-a;

run;

.enka

ta/eddy0onasani

((7


119/121

+e,t

Steps

Venkata

Reddy

-ona

sani

11@

What ne,t??


120/121

What ne,t??

Venkata

Reddy

-ona

sani

1

SAS &asi's

LearnAdvan'ed

SASprogramingte'hni0ues

SAS!eve"oper

Re.: 4heLitt"e SAS

&ook

Learn !ata

Ana"yti's

!ataAna"yst

Re.:(ra'ti'a"&usinessAna"yti'susing SAS


121/121

4hank you

Venkata

Reddy

-ona

sani

11

Stepbysteplearningsasbasicsv1 150218054117 Conversion Gate02

Documents