Definition

A directory set is a group of one or more directories used for storing all the sequential files created for a single data extract. A directory set definition lists the possible data volumes for a data extract and maps those data volumes to physical directories, as in the following table:

Use

When extracting data, the user must specify a file name and a directory set in which to place the extracted data.

When configuring DART, you must create at least one directory set to allow data extractions. In it, you must associate at least one data volume with at least one physical directory.

If you define only one directory set, then all data extracts will go into that directory set. This may result in different data extracts sharing the same directory or directories. This arrangement may be adequate if you move the extracted files to another media on a regular basis. You treat the directory set as a temporary holding place for extracted data from which users archive for long-term storage. DART users can therefore reuse the same directory set for different data extracts over time.

If several DART users will be extracting data at the same time, and you wish to provide them with separate areas to work with, then you should create several directory sets.

Structure

You define a directory set assigning each of the possible data volumes to a physical directory. You must specify enough directories to accommodate the largest amount of volumes that will ever be created by DART users. As discussed in the data volumes topic, each volume corresponds to a single physical data extract file. There can be multiple physical files for a single data extract.

When assigning a data volume to a physical directory, you must specify the name of an existing physical directory using the appropriate syntax for path names on the operating system of your application server.

The following table shows an example of such an assignment on a Windows NT system:

The following table shows an example of such an assignment on a UNIX system:

If you use the default file path syntax, the directory name in both operating systems must contain the ending slash (whether backslash or forward slash.) If you omit the ending slash, the system ignores the name of the last subdirectory, in effect using the directory above it to store the files. In the above examples, the system will use mydir , rather than mysubdir to store the files, if you omit the slash or backslash after mysubdir .

You can divide the volumes (which represent the data extract files) among one or more physical directories. You may, for example, specify that all of the volumes are stored in the same directory. You do so by entering the same directory name for all the possible volumes. Alternatively, you may specify that several of the volumes are stored in a single directory, or you may specify that every volume is stored in its own separate directory.

One of the factors affecting how you divide the volumes among the physical directories is the maximum file size you specify in the Display settings for data extraction screen.

If the directory size is limited to 2 Gigabyte on your operating system, and you specify a maximum file size of 1 Gigabyte, then make sure that no more than two volumes are stored in each physical directory.

The following table illustrates how you would assign volumes to directory sets in the above scenario (The example uses the Windows NT syntax.):.

If you determine that all data extracts created by the user will fit into a single file, then, at a minimum, you must define one volume and assign it to one physical directory.

See Also

See the file path syntax topic for a discussion of how the full path name is constructed.