SAS 'O' LOG

How to create a PDF ﬁle of a SAS job

2014-03-06T23:40:00.002-08:00

For creating a PDF file of a SAS job or SAS procedure, we need to bracket the section of code for which you want to create tables or/and graphs:

options orientation=landscape;
ods pdf file=’test.pdf’;
goptions device=pdf colors=(black) rotate=landscape;
...
\SAS\ procedures
...
ods pdf close;

Example:

options orientation=landscape;
ods pdf file=’output-to-Test.pdf’;
goptions device=pdf colors=(black) rotate=landscape;

ods graphics on;

SAS Interview questions and answers Part 1

2014-02-26T23:48:00.000-08:00

Do you have scheduled interview and want to know about the general questions asked in the SAS interview?
Here are some of the general questions asked in an interview with their answers.

When would you use select instead of IF statement?
A: Select is used over IF statement when you are using a condition to compare with several other conditions: Here is an example:

Data exam;
Set exam;
select (pass);
when Physics >60;
when math > 100;
when English = 50;
otherwise fail;
run;

What is the one statement to set the criteria of data that can be coded in any step?

A) Options statement.

What does ERRORS = 1 do in the OPTIONS statement?
A)If there an error in the data for a observation then ERROR variable will have a value of 1 else 0

Is there any difference between VAR A1 - A4 and VAR A1 -- A4?
A) No, there is no diff between VAR A1-A4 and VAR A1—A4. You can use single (-) or double (--) but if you use more than 2 (---) then it will thrown an error.

Have you seen the error message ""numeric values have been converted to character"? What does it mean?

A) It means SAS has automatically converted the numeric variable to character to perform the functions.

Why is a STOP statement needed for the POINT= option on a SET statement?

A) Because POINT= reads only the specified observations, SAS cannot detect an end-of-file condition as it would if the file were being read sequentially.

Can you control the number of observations read in by SAS?
A) Yes, by using FIRSTOBS and OBS option

Approximately what date is represented by the SAS date value of 730?
A) 31st December 1961

Does SAS 'Translate' (compile) or does it 'Interpret'? Explain.
A) Compile

What does the RUN statement do?
A)SAS editor will start compiling the data or proc step as soon as it encounters the run statement. SO if there are more than one data steps or a proc step followed by a data step then you can avoid using the run statement.

How to retrieve the first character of a variable in SAS

2014-02-25T23:44:00.004-08:00

If you need to create a variable which has only first character of the character variable, then there are 4 different ways with which you can do the same in SAS. Below each variable B1,B2,B3 and B4 contains Letter A, but the length is not the same for all the four variables.

data test;
 name='Alpha';
 B1 = substr(name,1,1);
 B2 = first(name);
 substr(B3,1,1) = name;
 length B4 $1;
 B4 = name; 
run;

Read data directly from websites in SAS

2014-02-25T23:39:00.004-08:00

Do you know you can read data from websites directly? You can use the URL Access Method in the filename statement in SAS.

filename seercode URL "http://www.abc/data.txt";
data siterecode;
infile seercode truncover;
input @1 bigline $char256.;
run;

How to clear library of data in SAS

2014-02-25T23:29:00.001-08:00

At times there is a need to clear the permanent library of datasets at one go. There is an option in SAS, PROC DATASETS provides a solution on the proc statement with a Kill Option.

Here is an example:

PROC DATASETS LIBRARY = TEMP KILL;
RUN;

Aliases in SAS

2014-02-25T23:24:00.002-08:00

I could choose to let “pp” be
aliased to “proc print; run;”
→
Should I want to add a print statement to
my code, I only need to type:
pp
<enter>
and SAS will substitute in the “proc print”
statement.

You can add as many as you want and can
edit them at a future date

SAS will save these for your next session in
your SAS profile

These work in Windows, but not sure about
UNIX OS

Automatic email from SAS after program completion

2013-10-14T02:29:00.002-07:00

If you want to sent an email as soon as your program completes, try:

sas program.sas ; echo program.sas is done | mail youruserid

If you don't want to regularly check the status of the program and wants to get notified to start your next job
If you only want an email if your program bugs out, try:

sas program.sas || echo your program didn't work | mail youruserid

You can write anything that you want in your email after echo.

Limit the size of dataset for test

2013-10-14T00:07:00.001-07:00

Sometime datasets are too large and to run the test on those datasets to check if the program is working fine, it takes a lot of time.

You can overcome this by OBS = option: Here is an example:

Example:
data sub85;
set lib.all85 (obs=1000);
proc glm data=sub85;
model icd90=female black state1-stat49;

Again, you can also use this as a system option:

options obs=1000 [other SAS system options here];

You can remove it once the program is running smoothly.

Order of execution in multiple dataset options?

2013-10-10T11:20:00.001-07:00

When multiple dataset options are applied to a single statement, then the order of execution matters. The rule of thumb is that they are executed according to the alphabetical order of their names. That is, a drop option is executed before a keep, and a keep before a rename.

Now, given also that the variable number is decided by the order of first-appearance in the data step, can you guess the output of the code snippet below?

data one;
  set sashelp.class(rename=(name=eman) keep=age sex name drop=age);
run;
 
/* check */
proc contents data=one varnum;
run;

How to check for the existence of a directory in SAS?

2013-10-09T23:20:00.003-07:00

Did you ever encountered with an issue where you need to check the existence of a directory with a SAS code. Here is the code which let you do the same.

If directory exists macro will return 1 else 0.

%macro DirExist(dir) ; 
   %LOCAL rc fileref return; 
   %let rc = %sysfunc(filename(fileref,&dir)) ; 
   %if %sysfunc(fexist(&fileref))  %then %let return=1;    
   %else %let return=0;
   &return
%mend DirExist;

/* Usage */
option noxwait; /* SAS 8.2 only */
%put %DirExist(C:\Documents and Settings\);
%put %DirExist(Test);

Aligning Character Values with a Put Function

2013-10-09T02:30:00.005-07:00

We often use LEFT and RIGHT SAS functions to align the character values. This can also be achieved by PUT function.

Let us see this with and example:

data temp;
 
alphabetw='  a  ';
chardate=put(alphabetw,$10. -l);
output ;
chardate=put(alphabetw,$10. -c);
output ;
chardate=put(alphabetw,$10. -r);
output ;
 
run;

proc print;run;
 
Obs    alphabetw  chardate
 
 1        a       a
 2        a           a
 3        a                a

How to label the variable created by PROC TRANSPOSE

2013-10-07T04:58:00.000-07:00

In Proc Transpose, you can tell SAS, how to NAME the new variable created from the result of Proc Transpose.

Also by using the the value of the variable from the original dataset, you can tell SAS to LABEL the variables.

Below is the example demonstrating the same.

  proc transpose data=test out=test2 prefix=visit;
    by subjid;
    var labres;
    id visit;
    idlabel visitl;
  run;

Based on the values in the VISIT variable, SAS will name the variables VISITx .. VISITy. And will ALSO label VISITx with the corresponding value from VISITL on the given record and label VISITy with the corresponding value from VISITL on the given record for all created variables.

SUBSTR to replace

2013-08-30T01:34:00.003-07:00

SUBSTR can also to used to replace values of a string. In the below example the variable ID, where the first two characters can replaced by XX, with the help of the SUBSTR function

LAST      NAME    ID
Barczak  George  CP436
Adams   Linda     LA543
Preiss    Gloria    AU942

To change the first two letters of the ID to "XX", use the SUBSTR function on the left side of the assignment statement:

substr(employee_id,1,2)='DA';

Result:

LAST       NAME     ID
Barczak   George   XX436
Adams    Linda      XX543
Preiss     Gloria     XX942

No Warning Messages

2013-08-26T22:02:00.002-07:00

When you write a code ensuring no warning messages are thrown in log but this get the error message below.

WARNING 32-169: The quoted string currently being processed has become
                more than 262 characters long.  You may have unbalanced 
                quotation marks.

If so, check out the QUOTELENMAX/NOQUOTELENMAX system option. This option is particularly appropriate when you have long text strings in, for example, the ODS HTML TEXT= statement.

Save Disk Space in SAS

2013-08-12T06:42:00.004-07:00

Save disk space and make your data sets easier to understand by inputting only the data you need. Drop variables you no longer need.

DATA mylib.yearly(DROP=Rain1-Rain12);
SET Old(DROP=Snow1-Snow12);
Total = SUM(of Rain1-Rain12);
programming statements
RUN;

Drop DO loop indexing variables.

data Mylib.NewCost (DROP=i);
set Mylib.Cost;
array Amt(100) Amt1-Amt100;
do i=1 TO 100;
Amt(i)=MAX(0,Amt(i));
end;
run;

Store numeric categorical data in character variables to save space.
length quest1-quest40 $ 1;

Increase EFFICIENCY of a SAS code

2013-08-12T06:41:00.002-07:00

THINK Before You Type. One of the best practices of a good programming is planning before you type. Try to keep the code as simple as possible. Use procedures more than datasteps. For
example, permanently save only the final data set. Use temporary data sets to store intermediate results.

Instead of this:

data Mylib.New;
set Mylib.Sept Mylib.Oct;
programming statements
run;
data Mylib.District;
set Mylib.All Mylib.New;
programming statements
run;

Use this:

data New;
set Mylib.Sept Mylib.Oct;
programming statements
run;
data Mylib.District;
set Mylib.All New;
programming statements
run;

Check for the existence of a dataset

2013-08-10T09:51:00.003-07:00

So often there are so many datasets in a library that it becomes impossible to look out for a particular dataset. To help find a dataset, we can use a simple DATA step function for checking the existence if the dataset. The function is EXIST. Here is the macro to achieve this.

 %MACRO DATA_EXIST(INPUT);
  %IF %SYSFUNC(EXIST(&INPUT)) %THEN %DO;
  %PUT DATA EXISTS;
  %END;
  %ELSE %DO;
  %PUT DATA DOES NOT EXISTS;
  %END;
 %MEND;
 
 %DATA_EXIST(SASUSER.TEST);

Further simplification can be done by:

%MACRO D_EXIST(INPUT);
 %SYSFUNC(EXIST(&INPUT))
%MEND D_EXIST;
The macro %D_EXIST could be called from within an expression.
 %IF %D_EXIST(SASHELP.CLASS) %THEN %DO;

The Essence of PROC APPEND

2013-08-10T06:14:00.004-07:00

The APPEND procedure (and APPEND statement of the DATASETS procedure) is an efficient method for concatenating observations from a smaller data set to a larger data set. The BASE= data set option is reserved for the larger of the two data sets with the DATA= option for the smaller data set. Essentially, the APPEND procedure avoids reading any observations in the BASE= data set by positioning the record pointer at the end of the BASE= data set. Each observation from the smaller data set is then applied one at a time to the end of the BASE= data set. In the first example, the BASE= data set identifies a larger data set called MASTER and the DATA= data set identifies the smaller TXN data set.

Code:

PROC APPEND

BASE=master

DATA=txn;

run;

When two or more data sets need to be concatenated, multiple APPEND procedures are issued. In the next example, two separate PROC APPEND steps are specified to concatenate the two smaller data sets (TXN1 and TXN2) at the end of the larger BASE= (MASTER) data set.

Code:

PROC APPEND

BASE=master

DATA=txn1;

RUN;

PROC APPEND

BASE=master

DATA=txn2;

RUN;

Forcing the Concatenation of Data Sets

When one or more variables in the input data set, represented by the DATA= option, are not present in the BASE= data set, or the variables in the input data set are not of the same type as the variables in the BASE= data set, an optional FORCE option can be specified on the PROC APPEND statement to force the concatenation of the data sets. Using the FORCE option also prevents an error from being generated in these situations. The next example illustrates the FORCE option being specified in the concatenation of the TXN data set to the MASTER data set.

Code:

PROC APPEND

BASE=master

DATA=txn

FORCE;

RUN;

Selecting Observations for Concatenation

By default, all the rows of an input data set are typically concatenated to the end of the base data set. But when only selected rows of an input data set are desired, an optional WHERE clause can be specified for the associated input data set as a data set option. This is accomplished by specifying an optional WHERE clause. The WHERE clause defines the logical conditions that control which observations will be selected from an input data set, and ultimately concatenated to the end of the base data set. In the next example, a WHERE clause is specified to select only ‘PG’ and ‘PG13’ movies from the input data set for concatenation to the end of the base data set.

Code:

PROC APPEND

BASE=master

DATA=txn (WHERE=(rating in ('PG', 'PG-13')));

RUN;

Designating the Master

It is not at all a given that BASE= should point to the longer of the two data sets and DATA= to the shorter. One might do the opposite and use the shorter as the base

§ because one has been established as the master and the other is merely a container for a subset

§ to preserve structural properties (such as variable order)

§ because the two are in different libraries and the concatenation belongs in one of those and not the other

§ to maintain BY group ordering

§ to preserve a name for the concatenation without the need to invoke PROC DATASETS to swap names

If none of these concerns arise, and one is indifferent about the order, and one knows which data set is the larger of the two, it does make sense to reduce I/O by making the larger data set the base.

Conclusion

The APPEND procedure is a powerful tool for SAS users to use for data set concatenation purposes. Users can also concatenate unlike data sets using the FORCE option on the PROC APPEND statement as well as select which observations are captured from the input data set using a WHERE clause and concatenated to the end of the base data set. Users are encouraged to explore the many capabilities the APPEND procedure (or the APPEND statement of the DATASETS procedure) has to offer.

Proc Datasets Tip

2013-08-10T06:14:00.001-07:00

Delete all labels in th.pnsmutaa

proc datasets library = th nolist;

modify pnsmutaa;

attrib _all_ label='';

quit;

Rename th.bstpns_org to th.bstpns_org1

proc datasets library = th nolist;

change stpns_org = bstpns_org1;

quit;

Delete all in a library

proc datasets library = work kill nolist;

quit;

Delete one specific dataset from a library

proc datasets library = work nolist;

delete tst1;

quit;

Copy all in library th to work

proc datasets nolist;

copy out=work in=th;

quit;

Copy th.rappformat to work.rappformat

proc datasets nolist;

copy out=work in = th;

select rappformat;

quit;

Creating SAS dataset using datalines gives error if written inside a macro

2013-08-10T06:13:00.001-07:00

The error is something like this:

ERROR: The macro X generated cards (data lines) for the DATA step which could cause incorrect results.

The DATA step and macro will stop executing.

The solution is to insert the values in a table using proc sql.

The sample code below demonstrates the same.

options nomlogic mprint nosymbolgen;

%macro insertData;

proc sql;

create table class

(Semester char(25) ,

Professor char(24),

Location char(12),

Cost num(5));

quit;

proc sql;

insert into class

(Semester,Professor,Location, Cost)

values('Fall','Dr. Armstrong','London',3000)

values('Spring','Dr. Armstrong','Prague', 3300)

values('Spring','Dr. Hannity','Beijing',3000)

values('Fall','Dr. Hannity','London',3000)

values('Spring','Dr. Smith','Prague',3300)

values('Fall','Dr. Smith','Los Angeles',3800)

;

quit;

%mend insertData;

%insertData;

Using the ANY and ALL Operators in Proc SQL

2013-08-10T06:12:00.002-07:00

The ANY and ALL operators are Proc SQL operators which can be used to compare the result of a query expression against either ANY or ALL of the results returned from another.

The following example demonstrates its use as part of a subquery comparing the annual sales of sports shoes from our 'Mozart' company with the top brands nationwide. The ANY operator is used initially to determine which of our company brands are outselling any one or more of the market leaders.

data mozart_shoes;

format id z3. name $char24.;

input id name $ & sales;

datalines;

001 Super-tread trainer 50000

002 X-Pro sports trainer 60000

003 Road runner deluxe 55000

004 Triathlon special 30000

run;

data top_brands;

format name $char24.;

input name $ & sales;

datalines;

Cougar 500 58000

Mike running pro 56000

Abibas Super-tread 53000

run;

proc sql;

select name, sales

from mozart_shoes

where sales gt any (select sales from top_brands);

quit;

This query returns the following results as expected where these two sports shoes have annual sales greater than at least one of the top brands.

name sales

__________________________________

X-Pro sports trainer 60000

Road runner deluxe 55000

Changing the operator to all shows us the one shoe out competing all of the top brands.

name sales

__________________________________

X-Pro sports trainer 60000

The NOT operator can also be used to good effect as part of the condition. For example, changing our code to the following:

proc sql;

select name, sales

from mozart_shoes

where sales not gt any (select sales from top_brands);

quit;

returns all of the shoes whose annual sales did not exceed any of the top brands.

name sales

__________________________________

Super-tread trainer 50000

Triathlon special 30000

LARGEST and SMALLEST Functions

2013-08-10T06:11:00.002-07:00

The MAX and MIN functions allow you to easily identify the maximum and minimum values from a selection of variables. However in order to find the second highest value or the bottom 3 values you will need to make use of the SAS 9 functions LARGEST and SMALLEST.

The following examples are based on the data listed below:

data balances;

input custid $ bal1 bal2 bal3 bal4 bal5 bal6;

datalines;

12345 911.00 863.30 896.98 1672.26 1675.20 602.42

12346 62.27 10.28 1377.79 1683.42 1290.92 553.13

12347 1996.54 74.68 253.94 374.02 691.90 593.75

12348 1494.83 1114.74 332.76 505.95 . 963.98

12349 336.61 295.78 .a 1965.40 592.76 .b

;

run;

Maximum and Minimum

The syntax for the maximum and minimum functions is as follows.

function(list of variables);

Largest and Smallest

The largest and smallest functions have similar syntax.

function(value, list);

Where value is the nth item you want (e.g. 2 for the second largest/smallest) and

list is a list of variables that will be processed.

The following code demonstrates the use of the MAX, MIN, LARGEST and SMALLEST functions.

data balances2;

set balances;

maxbal=max(of bal1-bal6);

minbal=min(of bal1-bal6);

large1=largest(1,of bal1-bal6);

large2=largest(2,of bal1-bal6);

small1=smallest(1,of bal1-bal6);

small2=smallest(2,of bal1-bal6);

small3=smallest(3,of bal1-bal6);

drop bal:;

run;

The following results are produced:

Max, Min, Smallest and Largest Functions

maxbal minbal large1 large2

1675.20 602.42 1675.20 1672.26

1683.42 10.28 1683.42 1377.79

1996.54 74.68 1996.54 691.90

1494.83 332.76 1494.83 1114.74

1965.40 295.78 1965.40 592.76

Max, Min, Smallest and Largest Functions

small1 small2 small3

602.42 863.30 896.98

10.28 62.27 553.13

74.68 253.94 374.02

332.76 505.95 963.98

295.78 336.61 592.76

Note that where the first argument in the largest and smallest functions is set to '1', the result is the same as the max and min functions. Also note that in this example when evaluating the minimum, the missing values have been ignored.

Handling of missing values

The max and min functions completely exclude missing data from the evaluation however the LARGEST and SMALLEST functions handle missing values in the data by placing any missing values present at the end of the sorted list of variables.

So the LARGEST function will return the maximum data value where a '1' is specified in the first argument, the second largest for a '2' and so on. Once all of the non missing data is referenced the missing values will be returned.

The SMALLEST function will return the lowest value when '1' is specified the second lowest for a '2' and so on. Again once all the non missing values have been referenced the missing values will be returned. The last non missing value in this case will be the highest data value.

The following example demonstrates how the largest and smallest functions deal with missing data values.

data balances3;

set balances;

large1=largest(1,of bal1-bal6);

large6=largest(6,of bal1-bal6);

small1=smallest(1,of bal1-bal6);

small6=smallest(6,of bal1-bal6);

drop bal:;

run;

Generates the following data:

Missing Values with the LARGEST and SMALLEST Functions

large1 large6 small1 small6

1675.20 602.42 602.42 1675.20

1683.42 10.28 10.28 1683.42

1996.54 74.68 74.68 1996.54

1494.83 . 332.76 .

1965.40 . 295.78 .

SAS FORUM

2013-08-08T00:30:00.003-07:00

SAS 'O' LOG - Discussion

Password protect your data

2013-08-07T23:11:00.003-07:00

Want to protect your SAS dataset with a password to protect it from other users. You can use the pw = option in the SAS datastep. Here is an example demonstrating the same. This will protect your data with read/write authorization.

DATA <EXAMPLE DATA> (ENCRYPT = YES PW = <PASSWORD>);
SET <EXAMPLE DATA>;
RUN;

If you lose your password, a lot of time and effort is needed to decrypt the file.

Extract the first letter of the character variable

2013-08-07T23:06:00.001-07:00

Have you ever needed to create a variable which is the first letter of the character variable? In the below example there are four ways to do that. X1, X2, X3 and X4 all contain letter P.

data one;
 name='Post';
 X1 = substr(name,1,1);
 X2 = first(name);
 substr(X3,1,1) = name;
 length X4 $1;
 X4 = name; 
run;