JPS62191923A - Data processor - Google Patents

Abstract

PURPOSE:To execute the hash retrieval by an optimized processing amount by changing the array of records at the same source address according to retriev al frequency. CONSTITUTION:An array modification means E modifies the array of the records with the same source address in a data file memory means A according to their retrieval frequency shown by a retrieval frequency memory means D. Accordingly, for retrieving a specific record, a hash retrieval means B can find it comparison times depending upon the retrieval frequency of the record. Namely, if the retrieval frequency of the record concerned is substantially higher than that of other record, it is retrieved earlier than the other. As a result the retrieval by the hash retrieval means B is optimized in general.

Description

[Detailed Description of the Invention] [Field of Industrial Application] This IJ relates to a data processing device for data files,
In particular, it relates to techniques for organizing data within data files for optimization of retrieval of specific records.

[Summary of the Invention] According to the present invention, in a data file data processing device, the arrangement of records with the same source address is changed according to the search frequency, so that hash search can be performed in an optimized process as a whole. became possible.

[Prior Art] A hash search method is known as a method for searching a data file storage means for a specific record. this is,
Within each record of the data file, the term [1 (key data)
By providing a field for search [1 (key data for requesting a specific record)] and using it as an address (source address) for the data file after performing a predetermined arithmetic process on the search term [1 (key data for requesting a specific record)], search terms [1 and It searches for records that have matching items.

In this type of system, when creating a data file, the same arithmetic processing as when searching is performed to obtain an address within the file, and a record 117 is placed at that address. If the file is large enough and the key data is well distributed, it is unlikely that the address will be in the east, but it stipulates an arithmetic method that generates hash codes with different values from a large number of arbitrary key data. Since it is impossible to do so, address distortion is actually unavoidable.

In order to deal with this problem of address clumps, in the conventional example, when a clump of addresses is detected, recalculation is performed based on that address, and an empty area is searched for the record. )
It is crowded.

[The dark problem to be solved by Part II] Therefore, when searching for a specific record from the data file created according to the conventional example described in Section 2, the number of comparisons required is +1 the number of lf operations. This number depends on the arrangement order of the records that generate the address (source address) of the first calculation process. For example, if there are 10 records that generate the same source address,
Assuming that the 10th record among them is the most frequently searched, a minimum of 10 comparisons are always required to find this record, which is extremely wasteful processing. As can be seen from this example, in the conventional technology, there is a problem in that the number of comparisons or search time, which represents the effort involved in searching, is determined by the order in which records that generate the same source address are arranged at the time of creation, regardless of the frequency of searches. Ta. Therefore, it is an object of the present invention to provide a data processing device that solves these problems and enables efficient searches.

[Means for Solving the Dark Problem] FIG. 1 is a block diagram showing the mechanism of the present invention. In the figure, A is a data file storage means for storing a data file consisting of a plurality of records, B is a hash search means including a source address generation means for retrieving a specific record from the data file storage means A, and C is a hash search means. is a source address storage means for storing the source address of each record constituting the data file, D is a search frequency storage means for each record, and E is the data file storage means A mentioned above.
Arrangement changing means for changing the arrangement of at least two records having the same source address in the database based on their search frequency.

[Sakukawa] To explain the present invention in detail, the arrangement change means E changes the data file storage means A to records having the same source address in accordance with the search frequency of those records indicated by the search frequency storage means. The arrangement of those records within is changed. Therefore, when searching for a specific record, the hash search means B can find the record with the number of comparisons depending on the search frequency of that record. That is, if the search frequency for that record is substantially higher than the search frequency for other records with the same source address, then the record will be searched earlier than the other records. As a result, overall,
The search by hash search means B is optimized.

[Example] An embodiment of the present invention will be described in detail below.

FIG. 1 is a block diagram when the present invention is applied to data processing such as ECR, PO3, etc., where 1 is a data file memory constituted by a RAM. Each record in the sales data file memory 1 has as data elements a source address, a search counter, key data (FLU code such as JAN code), unit price, and number of items sold.

Reference numeral 2 controls the entire system using the CPU, and in particular, for the purpose of the embodiment, it has a function to change the arrangement of records in the data file memory 1 according to the number of searches. The calculation unit 2-1 of the CPU 2 executes various calculations, and controls various peripheral devices via the control unit 2-2.

3 is a ROM that stores programs used by the CPU 2.
It is.

4 is the key human power department, and the operation mode of the sales data processing device (P is setting, REG is registration, X is inspection, Z is payment)
Mode selection key 4-1, numeric keypad 4- for selecting
2. It is provided with various function keys 4-3 and the like. This key power from the key power section is sent to the CPU 2 via the input system Wi 115.

6 is a barcode scanner that reads product barcodes (PLU code format such as JAN code), and is used in the registration mode and setting mode, and the information is sent to the CPU 2.
sent to and processed.

7 is a group of working registers whose writing/reading is controlled by the CPU 2; the A register stores key data (FLU code) manually input from the key input unit 4 or the barcode scanner 5; The unit price input from section 4 is stored in the C register as the source address (first hash code) generated by CPU 2 from the key data.
is stored, and the target address (the next address calculated by the CPU based on the current address) obtained by performing one operation is stored in the D register. Also, the E register stores a specific record from the sales data file memory. Before finding, the target address of the detected record with the same source address is stored.

Reference numeral 8 denotes a display section for displaying various data and messages, and an error lamp 8-1 lights up when a search is made but no message is found. This display section 8 is controlled by the CPU 2.
This is performed by a display driving section 9 which is operatively coupled to the display driver 9.

Reference numeral 10 denotes a printing unit that issues receipts and journals, and its control is performed via a printing control unit 11.

Reference numeral 12 denotes a money drawer, and opening control of the drawer is performed via the CPU 2.

Operation Next, the operation of the embodiment configured as above will be explained in detail, particularly the creation and retrieval of data files.

Creation of records in the data file is performed in P (setting) mode. In this mode, the operator inputs the PLU code (key data) of the product to be set via the key input section 4 or barcode scanner 6, and enters the necessary product information, here the unit price of the product. into the flow.

First, the CPU 2 executes the first address calculation using the key data (FLU code) read from the A register, as shown in step Sl. This is done by calculating the value of the HASH function using the key data as an argument. The address obtained in this way is C as the source address.
It is stored in the register and the D register.

In step S3, the data file memory 1 is addressed by the contents of the D register (source address in this case), and it is checked whether the location is empty. If the area is not an empty area, the next address is calculated from the current target address which is the contents of the D register (step S4), and the process returns to step S3 again, and this process is repeated until an empty area is found.

If an empty area is found, the key data in the A register and the unit price in the B register are set in the key data setting area and width side setting area of the empty area, respectively (step S5).
Furthermore, the current source address in the C register is set as the intra-record source address (step S6).Finally, the intra-record search counter is initialized to zero (step S7).

Next, we explain the operation at the time of product registration. Regarding the product purchased by customer 0, when the operator inputs the FLU code, that is, the key data of the product through the key manual unit 4 or the barcode scanner 6, as shown in FIG. Get into the flow.

First, the CPU 2 initializes the comparison address +iif times by clearing the E register, as shown in step Tl. Next, calculate the first address from the input key data and store the result in the C register and D register as the source address (steps T2, T3), and then use the manually entered key data and the current address in the D register. Compare the key data in the record in the second data file memory lh specified by the target address and check for match/mismatch (step T4, T5). Therefore, it is passed to the registration processing routine Tll, but if there is a mismatch, the process proceeds to step T6, and it is determined whether the search is completed by checking whether the record is in an empty area. end r
In this case, the corresponding h record is not in the file, so an error lamp is displayed (step TlB).If the process is not completed, the source address obtained from the manually entered key data and the source address in the record are read and matched. It is determined whether or not (step T8).

If they match, the previous comparison address is updated by taking out the current target address from the D register and storing it in the E register as the previous comparison address (step T9); if they do not match, or after updating the comparison address, step T10 Proceed to and calculate from the current (four-dimensional address) to the next address and return it to the D register (step Tl0)
.

Then, the process returns to step T4, and the same process as described above is executed for this C-calculated address.

To summarize the above processing, the CPU 2 first generates a source address from the input key data, and then sequentially generates the next address from the previous address, and the key data of the record at each generated address is The process is repeated until the key data matches the input key data, that is, until the target record is found. Therefore, with the target address of the last encountered record having the same source address entered the E register as the previous comparison address, it is confirmed in step T5 that the target record has been found ( (except if there is a match in step T5 in the first search).

Now, when the target record is found in the process of step T5, the normal registration processing routine Tll is executed by updating the sales data in that record, and the search counter in that record is increased by +1. do. Then, in step T13, the contents of the E register are read, and it is checked whether there was a previous comparison, that is, whether a record having the same source address was encountered. If so, the contents of the E register are the target address of the last record with the same source address encountered, so the search counter in the record at that address and the record searched this time The content of the search counter in the search counter is read out, and it is determined whether or not the number of searches for the record searched this time is greater than the number of searches for the last record encountered (step TI4,
T15), and if the number of searches for the last record encountered, that is, the record at the previous comparison address, becomes smaller than the number of searches for the record searched this time, if the search continues in the same search order, the search frequency will decrease. Higher records require longer search time than lower records, resulting in lower search efficiency overall.

Therefore, if the number of searches for the record searched this time is greater than the number of searches for the record found before the current record, both records are replaced (step T16), so that the relationship between the search times of the two is corrected. The reverse is true, and an overall more efficient search will be performed from now on.

If No in step T13 and NO in step T15,
After the processing in step 716, the process proceeds to the final step T17, where the target record number and record contents are displayed, thereby performing normal sales data display.

In addition, in the embodiment described in (B), during a search, the records are arranged according to the number of times the search is performed only between the records with the same source address found immediately before the target record is found. However, the arrangement of these records may be changed according to the search frequency among all records with the same source address found until the target record is found.

In addition, in the above embodiment, the calculation of the next address is performed using only the current target address as an argument.
The calculation may be performed using the added data as an argument, and even in this case, search optimization is achieved by changing the arrangement of records based on the search frequency.

The present invention can also be applied to a data structure in which only a group of records having the same source address are sequentially accessed each time a next address (NEXT ADDRESS) is generated.

[Effects of the Invention] As mentioned above, according to the present invention, records with the same source address are rearranged according to the search frequency, so a data file optimized for the usage status of the system can be created. Can be built.

[Brief explanation of drawings]

FIG. 1 is a block diagram of the present invention; FIG. 2 is a block diagram of an embodiment in which the present invention is applied to a data sales processing device;
Figure 3 is a flowchart for setting records in the Sell-L data file, and Figure 4 shows how to search for a specific record from the Sell-L data file, register the Sell-L data, and arrange the records according to the search frequency. It is a flowchart which shows the operation|movement of an Example to change. l...Sell, E data file memory, 2...
...CPU, 3...ROM, 4...
Input key 17... Working register group. Figure 1 Figure 3

Claims (1)

[Claims] Data file storage means; search frequency storage means for storing the search frequency of each record constituting the data file; and a source for generating a source address for searching records from the data file storage means. hash search means including an address generation means; source address storage means for storing the source address of each record; and an array of at least two records having the same source address in the data file storage means based on their search frequency. 1. A data processing device comprising: arrangement changing means for changing the arrangement.