The send or receive file for transfer with openFT may be an individual member of a PO or a PDSE data set. In this case, the same file contents are transferred as when a sequential file is transferred. In contrast to the transfer of an entire PO data set (see below), no directory information is transferred. In this case, the following specifications must be made for the file name in the NCOPY command:
fully qualified: 'data-set-name(member-name)'
partially qualified: data-set-name(member-name)
The data set name can contain one of the prefixes ":L:", ":O:" or ":E:". If it does not have a prefix or has the prefix ":L:", the name represents any data set (PO or PDSE). To access a specific PO member, you must use the prefix ":O:"; for a PDSE member, you must use the prefix ":E:".
You may specify *REPLACE-FILE, *NEW-FILE and *EXTEND-FILE for the WRITE-MODE operand. In this case, you are referring to the individual member. An existing PO or PDSE data set is always retained. If necessary, the PO or PDSE data set itself is also created.
You can specify the type of the data set (PO or PDSE) using the prefixes mentioned above (":O:" for PO, ":E:" for PDSE). If you do not specify a prefix, or you specify the prefix ":L:", a default value defined by the FT administrator is taken as the type for the newly created data set.
These specifications have the following meaning:
*REPLACE-FILE:
The member is written to the data set even if a member with the same name already exists.
For a PO data set applies: The new member is then located at the end of the PO data set; there may be a gap in place of the old member.*NEW-FILE:
Transfer takes place only if no member with the same name exists.
*EXTEND-FILE:
The contents of any existing member with the same name are copied. The transferred data is appended and the resulting member is written to the data set.
For a PO data set applies: The extended member is then located at the end of the PO data set; there may be a gap in place of the old member.
It is clear from this description that the transfer of individual members may result in gaps in the PO data set.
If a new PO or PDSE data set is created on receipt of a member, then the primary allocation corresponds approximately to twice the size of the send file and the secondary allocation to four times its size.
If members are written to an existing data set then the primary and secondary allocations remain unchanged.
If necessary, you must make sure that the allocation for a PO data set that is to be received is sufficiently large. If a large number of members are to be entered then sufficient directory blocks must be available. If members are frequently replaced or extended then a PO data set should be periodically compressed. No compression is required for PDSE data sets.
File transfer requests with read access to a member can be restarted an unlimited number of times. Requests that write to a member can also be restarted. However, when transferring PDSE data set members during restarted requests, openFT restarts at the beginning of the file. Requests that extend existing members are only permitted for PO data set members. Please also note that if a restart has to be performed during transfer of a PO data set member, this member is always appended to the PO data set. It is therefore advisable to compress this type of PO data set from time to time.
In the event that, on receipt of a PO or PDSE member, the data set must be created again, and this creation attempt fails, the file transfer will be aborted and a error message output.
Possible reasons for this failure include, among other things, disk bottleneck or a faulty specification for one of the SMS classes. If this happens, you should speak to your FT administrator.