JPS6024657A - Data storing method - Google Patents

Abstract

PURPOSE:To reduce a memory capacity by a simple control by providing an overflow area, writing a data which cannot be contained in a unit storage area, in said area, and writng its address as a pointer in the unit storage area. CONSTITUTION:A record storage area becomes a two byte unit on amemory M. Accordingly, in case of writing a data, whether its data is <=15-bit or not is checked, it is written in a storage area (m) in case of <=15-bit, and written in an overflow area OVF in case of >=15-bit. Also, a pointer is written in the storage area (m) of the memory M to write its data in. By executing such a write, when most of records stored in a magnetic disk, a magnetic tape, etc. are 2 bytes, most of them are contained in a group of a storage area of a 2 bytes, most of them are contained in a group of a storage area of a 2 byte unit of the memory M, and a memory capacity can be reduced.

Description

DETAILED DESCRIPTION OF THE INVENTION Technical Field of the Invention The present invention relates to a method of storing data in a memory with the aim of reducing the capacity of a memory in which data is expanded during data editing.

Prior art and problems Magnetic tape family rL (MT) or magnetic tape family (D
If you want to edit the data stored in K), store (deploy) the data in memory (main memory of the computer).
Then, processing such as rearranging the data in the memory according to rules such as order of magnitude is performed. For example, an employee database is configured by using a unit called a record that includes a plurality of items, and storing various information regarding the employee in each item of the record. If you have a database like this, for example, if you want to know the actual situation of salaries by age in the organization, you can create a series of records that include age and salary as items in the database, and arrange them in order of age. It is only necessary to perform statistical processing on their salaries, and memory is used to edit the data records. It is still important to check the usage frequency of a record for the purpose of determining the usefulness of the record, but for this purpose, it is necessary to set up a section 1 to record the number of times the record has been read. It is only necessary to read out and total up the information, and when doing so, the first layer of the memory is also opened.

By the way, if the data is 2 high I. or 4 high data, it is not necessarily constant, and is stored in the MT, DK, etc. in an undefined length state. However, when expanding to memory, it is treated as a constant length for ease of address assignment, etc. If the record length of the data header is 2 bytes and 4 by 1-, the longer one is taken and the record length is fixed at 4 bytes and the storage area is was secured.

However, in this case, only half of the storage area is used when storing 2-byte records, and records in data files are mostly 2-byte, and 4-byte records are rarely used, which is a noticeable waste.

OBJECTS OF THE INVENTION The present invention attempts to address these problems and makes it possible to reduce memory capacity without complicating control.

Structure of the Invention The present invention provides a method for storing variable-length data when expanding and editing the variable-length data on a memory. Write variable-length data sequentially to the unit storage area group \, and write long data that does not fit in the area in section 1.1) to the overflow area, and write its address to the unit storage area j ( where the data should be stored). The feature is that it is written in one area as a pointer, which will be explained next with reference to the implementation command.

Embodiment of the Invention In the present invention, as shown in FIG. 2, the record storage area in the memory MT is in units of 2 bytes. In other words, each address of the memory has a storage capacity of 1 by 1-, and in the conventional method, addresses θ ~ 3.4 ~ 7, 8 ~ 11 degrees, etc. are used to store a storage area of 4 bytes each. Using external storage device MT
Alternatively, each data record from the DK was stored, but in the present invention, each 4-hide storage area is divided into two and each data A, B, etc. is stored in a 2-byte storage area In. do. Data eight.

r3. If C,... is less than 2 bytes, storage area m
However, if it is 2 bytes or more, writing is impossible. Therefore, in the present invention, when writing, it is determined whether the data is 15 bits or less, and if it is less than 15 bits, it is written to the storage area m, and if it is 16 bits or more, 1 is written to the overflow area, and then the data is written. A pointer is stored in the storage area rn of the target memory M. In this example, data A, B, and C are 15 bits or less, and data D is 16 bits or more.Therefore, data A, B, and C were written to the storage area rn- of memory M, but data D was written to its pointer. DP-t-Writing.

As shown in the block P on the right side of the pointer, the 1-bit overflow display flag F and the GM of 5bit 71.
It consists of a total of 16 bits, including an L entry number N1 and an IO bit opaf 1 core entry number N2. Flag F
” indicates overflow. GML is overflow area 0
■F address table, GML entry number N
l indicates the address of table GML. The overflow area OVF has multiple pages PGl, PG2 . It consists of pages PGI, PG2. . . . and the overflow entry number N2 indicates the address within the page. Data D only pointer D'
It is written to the en1 edge of the page in the overflow area indicated by P. - For example, the number of entries in GML is 32, so page PG4. PG2. The number of... is 32, and the total number of entries in the overflow area is 32X1024=32768
, the storage capacity of each entry is 4 bytes.

In this way, if most of the records stored in DK and MT are 2 bytes, most of them can be stored in the 2-byte storage area group of memory M, which is much better than the conventional method. The required capacity can be halved.

In addition, even if there is data of 2 bytes or more, it is written to the overflow area, the write destination address is used as a pointer, and the J-do is stored in the storage area rn where it should be used (for example, in the above example) Then data D is data C
Since the record should be written in the area adjacent to the record, there is no support for reading the record. That is, data A, B, C,...
..., the access address should be 0.1, 2, .
You can proceed as follows, and this will result in data A.
,B,C.

is read out, and in the next D, the pointer DP is read out, and from the contents thereof, the storage area of the data D in the overflow area is accessed via the table GML, and the data D is read out.

As described in detail, the present invention is advantageous in that it is possible to reduce the required capacity of memory for storing (expanding) data during data processing, and of course, there is no problem in data processing. .

[Brief explanation of the drawing]

1'lil is a diagram for explaining a conventional method of developing data onto a memory, and FIG. 2 is a diagram for explaining the method of the present invention. In the drawing, M is a memory, A, B, C, . . . are variable length data, m is a unit storage area, OVF is an overflow area, and DP is a pointer. Applicant Fujitsu Co., Ltd. Attorney ± 11 Minoru Yanagi

Claims (1)

[Claims] In a method for storing variable-length data when expanding and editing the variable-length data on a memory, a storage area capable of storing up to a predetermined length of the variable-length data is used as a unit; A memory characterized by sequentially writing variable length data into a group of unit storage areas, writing long data that does not fit into the storage area into an overflow area, and writing the address as a pointer to the unit storage area where the data is to be written. How to store data on top.