Oracle^® Business Intelligence Disconnected Analytics Administration and Configuration Guide > Synchronizing Applications in Oracle BI Disconnected Analytics > Step 4: Setting Up for Preprocessed Synchronization >

Storing Data Set Files

With this version, the storage of data sets is more efficient. In previous versions, a set of iBots was executed for each user, regardless of whether there was commonality among data sets for different users. This version uses the query request key functionality of Oracle BI Presentation Services such that a sourcing query that generates the same data sets for two or more users will not be run multiple times. The query will be run only once and multiple users can access those data sets.

Configuration File Elements for Request Key Use

As in previous versions, store the result sets in the data directory for the application so that they will be available to users when they log on. With this version, the exact location differs depending on whether a request key is being used with caching in Oracle BI Presentation Services. Request key use is affected by the following elements within the <Disconnected> element in the instanceconfig.xml file on the Oracle BI Server:

• <UseRequestKey>: Enables or disables the use of the request key for caching in Oracle BI Presentation Services when the data is generated. The value is a boolean keyword of true (default) or false.
• <HashChars>: Specifies the characters to use when creating directories for preprocessed data. The value is an integer that defaults to 3. Use this element according to the following guidelines:
  • If a request key is not enabled, then this setting facilitates the naming of directories when user names are similar. For example, suppose you have two users with names of fred1 and fred2. You can set the <HashChars> element to a value of 3, which corresponds to "fre", so that the data for these two users is stored in files called fre/fred1/data.csv and fre/fred2/data.csv.
  • If a request key is enabled, then this setting facilitates the naming of directories and of files. For example, suppose that you set the <HashChars> element to a value of 3 and that the request key name is "abc123". The file is generated in a directory that uses the first three characters of the request key name, as shown in the following example:
    OracleBIData\Disconnected\myapp\data\abc\abc123.dat

Request Key Use and Naming of Directories and Files

The following list describes how request key use affects the naming of directories and files:

• If a request key is not enabled, then the data is generated in file that has a name that includes the following:
  • A directory that is named with the time stamp of when the data was generated.
  • A sub-directory that is named with the name of the user who requested the data.
  • A compressed file with a name that is created like so:
    • If the "file" attribute is specified in the <sourcingreport> element in the disconnected application configuration file, then the value of that attribute is used, followed by the suffix .DAT.
    • If the "file" attribute is not specified, then the value of the "name" attribute is used, followed by the suffix .DAT.

      The following lines show the format of this name:
      OracleBIData\Disconnected\app-name\data\timestamp\user-name\filename.dat

• If a request key is enabled, then the data is generated in file that has a name that includes the following:
  • A directory that is named with the time stamp of when the data was generated.
  • A file name that is the request key name with the suffix .dat.

    The following lines show the format of this name:
    OracleBIData\Disconnected\app-name\data\timestamp\filename.dat

For more information about where to store the result sets, see Step 1: Understanding the Directory Structure for Oracle BI Disconnected Analytics.