Using MSSQL TEMPDB in SAS

You can use the TEMPDB in Microsoft SQL-server through SAS by creating a ODBC-libname – like the libname below.

libname TMPLIB ODBC NOPROMPT="DRIVER=SQL Server; SERVER=<SERVERNAME>; DATABASE=TEMPDB; TRUSTED_CONNECTION=yes" schema=DBO CONNECTION=SHARED;

It’s important to provide the option CONNECTION=SHARED or else it will not work.

Through the libname it’s now possible to write and read from TEMPDB. The dataset has to have this syntax ‘#<DATASETNAME>’n e.g. like below ‘#temp’n

data tmplib.'#temp'n;
set sashelp.class;
run;

Be aware, that you are not able to view this new table through the Display Manager in SAS. When Microsoft SQL-server names the table, it makes the table name longer than SAS is able to display in the Display Manager.
You’re able to verify that the table do exist through SSMS (SQL Server Management Studio), or you can verify it’s existence by reading it back to SAS by using the code below.

proc sql;
create table temp as
select *
from tmplib.'#temp'n
;
quit;

If you want to use MSSQL-server temp-tables in Pass-Through SQL in SAS, then you need to use the libname option dbmstemp=yes.
Using this option will make it possible to execute the code below and force SAS to use the MSSQL-server to process the SQL-code in the pass-through SQL. If the option dbmstemp=yes is not used, then SAS will pull the data from the MSSQL-server back to be executed locally on the SAS-installation. It works with the below driver.

libname tmplib odbc noprompt="driver=odbc driver 11 for sql server; server=<servername>; database=tempdb; trusted_connection=yes" schema=dbo connection=shared dbmstemp=yes;

data tmplib.<SASTEMP-tablename>;
set <tablename>;
run;

proc sql noprint;
connect to odbc (noprompt="driver=odbc driver 11 for sql server; server=<servername>; trusted_connection=yes");
create table <tablename> as
select * from connection to odbc
(
select a.*
from <tablename> a
inner join tempdb.##<SASTEMP-tablename> b on a.<variable> = b.<variable>
);
disconnect from odbc;
quit;

Further information can be found in this link: http://support.sas.com/documentation/cdl/en/acreldb/63647/HTML/
default/viewer.htm#a002677192.htm

 

Extending the Display Manager in SAS

The commands below can be used to expand the Display Manager in SAS.

Command Description
vt &syslast.; Add to keys, e.g. F5. This opens a Viewtable with the latest run dataset/view.
next viewtable:; end; Add to keys, e.g. F9. Then F9 closes the last used table (that is open). Can close all views with multiple F9’s.
odsresults; select all; clear; wpgm; Add to keys, e.g. SHIFT F1. This key bind clears all SAS “results” and returns to the program editor.
log; clear; wpgm; Add to keys, e.g. F4. Clears log without having to highlight the log window. Returns to the program editor.

 

SAS “behind the scenes”

Below is a description of very useful options in SAS, if you want a look “behind the scenes” and see what SAS actually does when processing data.

options fullstimer sastrace=(,,,d) sastraceloc=saslog mprint source2 nostsuffix;
fullstimer The SAS System provides the FULLSTIMER option to collect performance statistics on each SAS step, and for the job as a whole and place them in the SAS log. It is important to note that the FULLSTIMER measures only give you a snapshot view of performance at the step and job level.
sastrace=(,,,d) Generates trace information from a DBMS engine.

‘,,,d’ specifies that all SQL statements that are sent to the DBMS are sent to the log. Here are the applicable statements:
SELECT
DELETE
CREATE
SYSTEM
CATALOG
DROP
COMMIT
INSERT
ROLLBACK
UPDATE
For engines that do not generate SQL statements, API calls and all parameters are sent to the log.
sastraceloc=saslog Prints SASTRACE information to a specified location.
In this case the log in SAS.
mprint Specifies whether SAS statements generated by macro execution are traced for debugging.
source2 Specifies whether SAS writes secondary source statements from included files to the SAS log.

SOURCE2 specifies to write to the SAS log secondary source statements from files that have been included by %INCLUDE statements.
nostsuffix The NOSTSUFFIX system option suppresses printing or display of trailing SASTRACE information and makes the SASTRACE log easier to read.

Getting first megabytes of large file

You can use the DOS-command TYPE and MORE to get the first megabytes or gigabytes of a large file e.g. of TXT, XML, CSV etc.
This is done by piping the result into another file, there by getting only the first part of the file piped into the new file.

The DOS-command below will start piping the large file into the smaller new file.

type LARGE_SOURCE_FILE | more > SMALL_DESTINATION_FILE

You are not able to see the result from the DOS-command. You need to press the SPACE-key on your keyboard, because every time you press the SPACE-key a new page of the file will be showed and thereby piped into the SMALL_DESTINATION_FILE.
When you think you have done this enough times, then you have to press CTRL-C on your keyboard, this will terminate the type-command and you will have the SMALL_DESTINATION_FILE containing data from the LARGE_SOURCE_FILE.

Be aware that you will get the — more — pipe text in the file!

SQL-statement to find lowest observation in dataset

The SAS SQL-statement below will create a dataset containing all information for the observation with the lowest age in the dataset SASHELP.CLASS.

proc sql;
     create table CLASS as
           select     *
           from SASHELP.CLASS
           group by AGE
           having AGE=min(AGE)
     ;
quit;

The syntax for the SAS SQL-statement is showed below.

proc sql;
     create table <DESTINATION TABLE> as
           select     *
           from <SOURCE TABLE>
           group by <COLUMN TO SEARCH>
           having <COLUMN TO SEARCH>=<FUNCTION>(<COLUMN TO SEARCH>)
     ;
quit;

As you see the <FUNCTION> doesn’t have to be min (minimum), it can be any function working on the type of <COLUMN TO SEARCH> – numeric or char.