07.08.2013 Views

SAS 9.1.3 Intelligence Platform: System Administration Guide

SAS 9.1.3 Intelligence Platform: System Administration Guide

SAS 9.1.3 Intelligence Platform: System Administration Guide

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

Using Configuration Settings to Optimize Metadata Server Performance R Repairing Problematic Data Sets 107<br />

Avoiding Changes to the Repository That Adversely Affect Performance<br />

The <strong>SAS</strong> Metadata Server optimizes its performance by relying on the original<br />

internal structure of repository data sets. Therefore, you should not use <strong>SAS</strong> options<br />

that affect the internal structure of these data sets, because such changes will<br />

adversely affect performance. For example, do not set the data set options COMPRESS<br />

and REUSE to Yes in order to compress observations and to reuse compressed space in<br />

compressed data sets. In addition, using a procedure like PROC SORT can cause<br />

problems. The problems result from the procedure’s changing the order in which<br />

observations are stored in the data sets.<br />

Determining Whether the COMPRESS or REUSE Attribute Has Been Set<br />

It is possible that the installer at your site set the COMPRESS and REUSE options<br />

on your repository data sets when he or she set up your system. When the installer<br />

creates your foundation repository by using <strong>SAS</strong> Management Console, he or she has<br />

the chance to specify data set options such as COMPRESS=Yes and REUSE=Yes, and<br />

may have done so in an attempt to conserve disk space. In addition, these options can<br />

be set inadvertently if they are specified as global options in a system command line, a<br />

configuration file, or an autoexec file when you start the metadata server or run the<br />

%OMABAKUP macro.<br />

To determine whether COMPRESS or REUSE has been set on a repository, run PROC<br />

CONTENTS against the repository directory. For example, you might run this program:<br />

libname reposdir ’C:\<strong>SAS</strong>\EntBIServer\Lev1\<strong>SAS</strong>Main\MetadataServer\<br />

MetadataRepositories\Foundation’;<br />

proc contents data=reposdir._all_;<br />

run;<br />

The output of this procedure will contain a list of variables and attributes for each data<br />

set processed. If the list for a data set indicates that the COMPRESS or REUSE option<br />

has been set, you will know that the internal structure of that data set has been altered.<br />

Repairing Problematic Data Sets<br />

To repair repository data sets that have been compressed, perform these steps:<br />

1 Create a temporary repository directory to hold the uncompressed data sets.<br />

2 Use PROC COPY or a DATA step to create an uncompressed version of the data<br />

sets in the temporary location. For example, you can use code similar to the<br />

following:<br />

/* set the correct system options for the <strong>SAS</strong> session and assign librefs<br />

to the original and temporary repository directories */<br />

options compress=no reuse=no;<br />

libname reposdir ’C:\<strong>SAS</strong>\EntBIServer\Lev1\<strong>SAS</strong>Main\MetadataServer\<br />

MetadataRepositories\Foundation’;<br />

libname tempdir ’C:\<strong>SAS</strong>\EntBIServer\Lev1\<strong>SAS</strong>Main\MetadataServer\<br />

MetadataRepositories\FoundationTemp’ compress=no;<br />

/* list the contents of the repository directory */<br />

proc contents data=reposdir._all_;<br />

run;

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!