Description
This test checks if the DB2 for Linux, UNIX and Windows data rowcompression feature is enabled for fact tables of InfoCubes andaggregates and for DataStore object tables. Data row compression isavailable in version 9 of DB2 for Linux, UNIX, and Windows (in thefollowing referred to as DB2 LUW).
Data row compression uses a static dictionary-based compressionalgorithm. It allows the replacement of repeating patterns that spanmultiple column values within a row containing shorter symbol strings.With data row compression, you can save disk space, reduce disk I/O andreduce data access times.
The test checks the following information that is related to DB2 LUWdata row compression for each fact table of an InfoCube and itsaggregates or DataStore object tables:

• Is row compression enabled?


• What is the size of the data object of the table?


• Does the table have a compression dictionary?


• If the table has a compression dictionary, is the size in KB or bytes
checked?
The test returns status "red" if there are tables with the followingfeatures:

• Data row compression is enabled


• The size of the data object exceeds 1 MB on each database partition
where the table resides

• The table does not have a compression dictionary

Otherwise, the status is "green".

Repairs
For tables with status "red", the repair function schedules one of thefollowing jobs:


• An offline reorganization job

Offline reorganization creates and stores a compression dictionary andcompresses all existing rows in the table.

• A job that calls the DB2 LUW INSPECT command.

The INSPECT command uses a sampling technology to estimate the amount ofdisk space that can be saved and to create and store a compressiondictionary. Existing data is not compressed, but data that is insertedor updated in the future is compressed. INSPECT is much faster thanoffline reorganization because it only processes a sample of the data.
You are prompted if you want to run offline reorganization or INSPECT.
The RSADMIN parameter DB6_ROW_COMPRESSION controls whether InfoCube andaggregate fact tables, DataStore object tables and PSA tables arecreated with data row compression enabled.
Data rows can only be compressed if you have created a compressiondictionary. You need to load at least a small amount of data (1 MB perdatabase partition on which the table resides) before a compressiondictionary can be created. If you create the compression dictionary witha larger data volume, the compression factor might be higher.
The standard procedure in SAP NetWeaver 2004s is as follows:

• By default, data row compression is disabled. If you want InfoCube and
aggregate fact tables, DataStore object and PSA tables to be createdwith data row compression enabled, you must set the RSADMIN parameterDB6_ROW_COMPRESSION to YES. For releases higher than SAPNetWeaver 2004s, the default value for DB6_ROW_COMPRESSION isYES.

• You load one or more data requests into the tables (at least 1 MB of
data multiplied by the number of database partitions where the tableresides).

• To create a compression dictionary for tables that do not have one yet,
you run this RSRV test and choose Repair to schedule an offlinereorganization or an INSPECT.
The data rows of very narrow tables might be too small to be compressed.In this case, the job log of the offline reorganization contains the SQLwarning -2220 "The compression dictionary was not built for one or moredata objects". The INSPECT job log contains the message "No dictionarycan be built for DAT object as no eligible rows were found". You cannotuse data row compression for these tables. We recommend that you disabledata row compression manually for them at the database level with theALTER TABLE statement.