JPS6349255B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION Technical Field of the Invention The present invention relates to a member replacement method for a partitioned data set, and particularly allows member replacement even when there is no extra free space.

Technical Background Computing centers have a library of several programs. All of these are partitioned and stored in the data area of memory, and the address of each partition (member) is recorded in the directory section of the memory to make it easier to search. This includes the operating system (OS) and compiler provided by the manufacturer, as well as the user's own business programs. The latter business programs may number in the hundreds. Data set members are updated when they become old, but usually there are 10 members.
Approximately 100 pieces will be replaced at one time. Replacement is performed by writing a new member into a free area of memory, but if there is no more free space for writing, the update will fail and you will have to allocate space and start over. Interruption of this replacement work must be avoided as much as possible, as it will halt business operations.

Prior Art and Problems Conventionally, in order to prevent the above-mentioned replacement operation from failing, a sufficient free space is secured in advance in the partitioned data set before replacement is performed. This will be explained with reference to FIG. This figure is an explanatory diagram of one partitioned data set 10, where 11 is a directory and 12 is a data area. directory 1
1 contains member A stored in data area 12.
~Member names of C~ are registered, and each member name is accompanied by the start address of the corresponding member. If there is a free space 13 in the data area 12, other members can be stored there. For example, members D, E, F
(not shown) is stored in the free area 13, the member name, and the barter area 12 of the member are stored.
The start address of is registered in the directory 11. To replace (update) members, use directory 11.
This differs from storing a new member in that it uses the same member name within. Member replacement using the conventional method,
Taking member A as an example, first store member A' (new member A) in the free area 13, and switch the address attached to the member name in the directory 11 from the address of member A in the data area 12 to the address of A'. , this completes the update. Member A, which has been updated and has become an old member that will never be read again, remains as it is. In this way, in the conventional method, when updating a member of a dataset, a new member is first written to the dataset, so in order to replace members A to C at the same time, the same amount or more of free space 13 must be secured. There must be. If members A to C occupy more than half of the data area 12, it is impossible to replace them. In this way, in the conventional method, when all members are updated, data can be stored in less than half of the data area 12, which has the drawback of extremely low usage efficiency. Moreover, some members have been updated and are no longer needed, so the usage efficiency is even lower when considered in terms of effective data.

Purpose of the Invention The present invention aims to improve the above points, and when updating a member, an empty area is created by deleting and condensing the member to be replaced, and the empty area is filled with the empty area. By writing new members, it is possible to update members without using extra free space.

Structure of the Invention The present invention provides a member replacement method for a partitioned data set in which a plurality of members are partitioned into data areas and the names and addresses of these members are registered in a directory. A process of deleting members of a second partitioned data set having the same member name as a member name of the first partitioned data set, a process of condensing remaining members in the second partitioned data set, and a process of condensing members remaining in the first partitioned data set. This method is characterized in that the members of the set are sequentially copied into the free data area of the second partitioned data set, and this will be explained in detail below with reference to the illustrated embodiment.

Embodiment of the invention FIG. 2 is an explanatory diagram showing an embodiment of the present invention.
shows a partitioned data set 20 in which members A' to C' are stored in its data area 22. In this example, members A' to C' of this partitioned data set 20 are
Copy it to the partitioned data set 10 of
. In addition to members A to C, members D to F are also stored in the segmented data set 10 of b, and there is not enough free space 13 left to store members A' to C'. Therefore, in the conventional method, updating is not possible, but in the present invention, this is possible in the following manner. That is, the member names 〓〓 to 〓〓 registered in the directory 21 of the partitioned data set 20 are sequentially read (GETD
(macro), this erases the partitioned data set 10 (STOW macro). This is directory 2
Continue until the end of 1. c is a state diagram of the partitioned data set 10 in which members A to C have been deleted, and in addition to the existing free area 13, there is also a free area 1 due to deletion.
4 occurs. In order to connect this empty area 14 to the empty area 13, condensation processing is performed to connect members D to F.
Repack from the beginning of data area 12 (JSECOPY program). With a large free area created in this way, the members of the partitioned data set 20 are
A' to C' are copied to the corresponding free area of the partitioned data set 10. (JSECOPY program). This is state d. The conventional system also has a condensation function, which is used to remove unnecessary members when there is no space left in a segmented data set or when there are unnecessary members. Here, we will use the condensation function.

In the member replacement described above, if members A' to C' have the same amount of data or less as members A to C, free space 1 will be used.
It is possible without 3. Note that if a part of the member name corresponding to the member names 〓〓~〓〓 on the partitioned dataset 20 side, for example, does not exist on the partitioned dataset 10 side, it cannot be deleted. This can be safely ignored if the request is only to update partitioned dataset 10, and if the request is to write partitioned dataset 20 to partitioned dataset 10, member A'
All you have to do is write a new file to the directory and register the member name and address in the directory.

Effects of the Invention As described above, according to the present invention, there is an advantage that member replacement of a partitioned data set can be performed without using extra free space.

[Brief explanation of drawings]

FIG. 1 is an explanatory diagram of a conventional member replacement method for a partitioned data set, and FIG. 2 is an explanatory diagram showing an embodiment of the present invention. In the figure, 10 and 20 are partitioned datasets, 11,
21 is a directory, 12 and 22 are data areas,
A to C are members to be replaced, ~ is the member name, A' to C' are members to be copied, and D to F are remaining members.

Claims (1)

[Claims] 1 In the member replacement method for a partitioned data set in which multiple members are stored in data areas and their member names and addresses are registered in a directory, the first
A process of deleting a member of a second partitioned data set having the same member name as a member name of the partitioned data set of the first partitioned data set, a process of condensing the members remaining in the second partitioned data set, and a process of condensing the members remaining in the second partitioned data set. 1. A method for replacing members of a partitioned data set, characterized in that members of the set are sequentially copied into free data areas of the second partitioned data set.