When accessed sequentially, through BSAM or QSAM, the PDSE directory appears to be constructed of 256-byte blocks containing sequentially ordered entries. The PDSE directory looks like a PDS directory even though its internal structure and block size are different. PDSE directory entries vary in length. Each directory entry contains the member name or an alias, the starting location of the member within the data set and optionally user data. The directory entries are arranged by name in alphanumeric collating sequence.

You can use BSAM or QSAM to read the directory sequentially. The directory is searched and maintained by the BLDL, DESERV, FIND, and STOW macros. If you use BSAM or QSAM to read the directory of a PDSE which contains program objects with names longer than 8 bytes, directory entries for these names will not be returned. If you need to be able to view these names, you must use the DESERV FUNC=GET_ALL interface instead of BSAM or QSAM. Similarly, the BLDL, FIND, and STOW macro interfaces allow specification of only 8-byte member names. These are analogous DESERV functions for each of these interfaces to allow for processing names greater than 8 bytes. See PDS Directory for a description of the fields in a PDSE directory entry.

The PDSE directory is indexed, permitting more direct searches for members. Hardware-defined keys are not used to search for members. Instead, the name and the relative track address of a member are used as keys to search for members. The TTRs in the directory can change if you move the PDSE, since for PDSE members the TTRs are not relative track and record numbers but rather pseudo randomly generated aliases for the PDSE member. These TTRs may sometimes be referred to as Member Locator Tokens (MLTs).

The limit for the number of members in a PDSE directory is 522,239. The PDSE directory is expandable; you can keep adding entries up to the directory's size limit or until the data set runs out of space. The system uses the space it needs for the directory entries from storage available to the data set.

For a PDS, the size of the directory is determined when the data set is initially allocated. There can be fewer members in the data set than the directory can contain, but when the preallocated directory space is full, the PDS must be copied to a new data set before new members can be added.