JPH02190971A - Index updating system - Google Patents

Abstract

PURPOSE:To reduce the overhead of an output to an intermediate file and to shorten the updating processing time of an index by collectively sorting the index updating information stored in the intermediate file and then updating an index part in accordance with the sorted contents. CONSTITUTION:At the time of updating a file 5, a data part 6 is updated based upon a series of updating data and a conversion means 1 convert updating information including key values extracted about plural index items included in respective updating data into index updating information collectively consisting of a sortable format character string in the order appropriate for the updating of respective indexes in respective index parts 7-1 to 7-n and stores the converted updating information in the intermediate file 3. After ending the updating of the data part 6, the index updating information stored in the file 3 is collectively sorted to form the index updating information arranged in the order appropriate for the updating of respective index parts 7-1 to 7-n and updating is executed in accordance with the contents. Consequently, the overhead of the output to the file 3 can be reduced and the updating processing time of the indexes can be shortened.

Description

Detailed Description of the Invention [Industrial Application Field] The present invention relates to updating of an index section in an information processing system in which a file including a data section and a plurality of index sections is stored in a randomly accessible storage device. In particular, the present invention relates to an index update method in which the data section is updated according to a series of update data and then the index section is updated all at once.

[Conventional technology]

There are various file organizations that can search for data with a predetermined item value from a file that stores a large amount of data including multiple items, but one of them is the data section that stores data. Separately, index data that has the value of each item of each data as a key value, and information about the location of the data that has the item of that value, such as the address of the data itself and the management number for shipping that address, is used as an index item. There is a file with the organization method stored in the daily index section.

In such a file, when acquiring data where the value of an item is α, the index section corresponding to that item is searched,
When index data having a key value of α is found, the corresponding data is acquired from the data section according to the location information in the index data.

By the way, when updating this type of file according to a series of update data given from the outside, it is necessary to update the data section and the index section at the same time, and this update was conventionally done as follows. was. That is, each time one piece of data in the data section is updated according to one piece of update data, the index update information for each index item is extracted from the updated data and output to an intermediate file provided for each index item. After the data section has been updated with update data, the multiple index update information stored in the intermediate file is sorted in an order suitable for update processing, such as sorting in the order of key values according to the definition attributes of the index items. ,
The process of updating the corresponding index section according to the sorted contents was performed for each intermediate file, that is, for each index item.

[Problem to be solved by the invention]

The conventional index update method described above updates the index asynchronously with the data update, sorts the multiple index update information in an order suitable for the update processing of the corresponding index section, and then updates the corresponding index section all at once. , the processing speed can be increased compared to a method in which the index section is updated sequentially every time one piece of data is updated. However, in order to obtain index update information arranged in the order in which each index section was updated, an intermediate file is created for each index item, and the index update information for each item extracted for one update data is stored in each intermediate file. By outputting to a file, only the index update information related to one type of index item is stored in one intermediate file, and then it is sorted for each intermediate file, so the process of processing one update data requires output to multiple intermediate files. Therefore, when a plurality of intermediate files exist on the same storage medium such as a magnetic disk, the overhead due to seek time becomes large, and the overall update time increases.

The present invention has solved these conventional problems, and its purpose is to provide an index update method that can reduce the overhead of outputting to intermediate files and shorten the index update processing time.

[Means to solve the problem]

In order to achieve the above object, the present invention includes a data section and a plurality of index sections that store index data for each index item, including key values of index items of data stored in this data section and location information of the corresponding data. Information processing that stores a file with a randomly accessible storage device in a randomly accessible storage device, updates the data section based on a series of update data, and then updates the index section related to the current series of updates all at once. In the index update method of the system, update information including key values extracted for multiple index items included in each update data can be sorted all at once in an order suitable for updating the index for each index section. The system includes a conversion means for converting into index update information consisting of columns, and an intermediate file for storing the index update information generated by the conversion means, and the index update information stored in the intermediate file is sorted all at once in the above order. After that, the index section is updated according to the sorted contents.

[Effect]

In the index update method of the present invention, when updating a file,
At the same time as the data section is updated based on a series of update data, the conversion means updates the index of each index section with update information including key values extracted for multiple index items included in each update data. The index update information is converted into index update information consisting of strings in a format that can be sorted all at once in a suitable order and stored in an intermediate file, and after the data section has been updated based on a series of update data, it is stored in the intermediate file. The index update information is sorted all at once to generate index update information arranged in an order suitable for updating each index section, and the index sections are updated according to the contents.

〔Example〕

Next, embodiments of the present invention will be described in detail with reference to the drawings.

FIG. 2 is a block diagram of main parts showing an example of an information processing system to which the index update method of the present invention is applied. File (intermediate file) 3
and an update data file 4.

The storage device 2 stores a mask file 5 to be updated. This mask file 5 has a data section 6 and multiple index sections for each index item? -1 to 7-n. The data section 6 stores a plurality of data. In addition, each index section 7-1 to 7-n stores index data that has the value of a corresponding item in the data stored in the data section 6 as a key value and also has information regarding the location of the data. . The data address itself, a data management number that is unique within the mask file 5, etc. are used as information regarding the location of the data, and in this embodiment, the data management number is used. Note that the address of each data is managed in an area (not shown) in the mask file 5 in association with a data management number, and the actual address is obtained from that area, but this is not directly related to the present invention. Details have been omitted.

FIG. 3 shows an example of the format of data in the data section 6 of the mask file 5. As shown in the figure, one data includes a plurality of index items fl-f5. Note that one piece of data is stored as one data record. The index items f1 to f5 illustrated here are as follows.

- Index item r1 Item identification - C1, Definition attribute = Character type 1 digit number = 12 digits (characters) - Index item f2 Item identification - 02, Definition attribute - Unpack format % expression % Item identification - 03, Definition attribute - Unpack format decimal number 1
Number of digits - 7 digits / Index item "4 Item identification = 04. Definition attribute = Pack format 10% formula % Item identification = 05. Definition attribute - Back format Decimal number 5 digits Number 11 digits Referring to Figure 2 again, update Data file 4 contains
Update data for the mask file 5, that is, data regarding data to be added, data to be deleted, and data to be changed is stored.

Taking the case of addition, which is one of the update modes, as an example, the update data file 4 stores additional data D1 to Di, . . . having a format as shown in FIG. The additional data Di illustrated here includes index items rl, f2 . The values of [3, [4, f5] are lined up. Note that one piece of additional data is stored as one additional record.

The processing device 1 in FIG. 2 is a device that updates the mask file 5 stored in the storage device 2 based on a series of update data stored in the update data file 4, and updates the data section 6. Later, an asynchronous index update is performed to update the necessary index sections all at once. In addition, the update information including the key values extracted for multiple index items included in the update data stored in the update data file 4 can be sorted all at once in an order suitable for updating the index for each index section. function to convert index update information consisting of character strings, function to output index update information generated by this function to work file 3, function to collectively sort index update information stored in work file 3, etc. To achieve this, processes 30-312 as shown in the flowchart of FIG. 1 are executed.

Next, we will explain the operation of this embodiment by taking as an example the process of adding additional data shown in FIG. 4 to the mask file 5 having the data section 6 in the data format shown in FIG. I will explain along.

When instructed to update the mask file 5 in the storage device 2 with the update data file 4, the processing device 1 starts the process shown in FIG.

(0) Initial setting processing SO Prepares the necessary operating environment such as generating work file 3 for sorting, opening mask file 5, etc.
; internal variable) -1.

(1) Update data import processing Sl Import additional data D1 from the beginning of the update data file 4, and use a data management number Rx that is unique within the mask file 5.
Assign.

(2) Data section update processing S2 Processes such as inserting additional data Di into the data section 6 of the mask file 5 are performed, and j (j; internal variable) is set to -1.

(3) Index update information extraction process S3 Update information diJ corresponding to the j-th index item fj in the additional data Di is extracted.

(4) Item attribute determination process S4 Determine the definition attribute of index item fj, and if it is character type, character type key value conversion process S5. If it is an unpacked decimal number, unpacked decimal key value conversion processing S6. Pack format 10
If it is a base number, control is passed to bank format decimal number key value conversion processing S7.

(5) Character type key value conversion processing S5 Leave the key value of update information d1j as is, and change the item identification CI
, the index update information kij is added with the data management number Rx and the value "1" corresponding to "addition" as the addition/deletion identification.
is created, and control is passed to index update information output processing S8.

FIG. 5(a) shows the index item f of the additional data Di in FIG.
Key value ABCDEFGH of update information dil corresponding to 1
An example in which index update information kll is created from ΔΔΔΔ is shown.

(6) Unpacked decimal key value conversion processing S Back up the key value of update information diJ, add the maximum value that can be taken based on the item definition to maintain the magnitude relationship of the key value as a character string, and convert the item identification Cj and data A management number Rx is added, and a value r corresponding to "addition" is added as an addition/deletion identification.
L" is added to create index update information kij, and control is passed to index update information output processing S8.

FIGS. 5(B) and 5(C) show the index item f2. of the additional data D+ in FIG. [Update information corresponding to 3 di2. Convert di3 key values 1234567 and -9876 to the maximum value 9999999999.9999 that can be taken according to the item definition.
Add 999 to 10001234566.999012
3 as index update information ki2. An example of creating +3 is shown below.

(7) Packed format decimal key value conversion processing 87 Add the maximum value possible in the item definition to the key value of the update information dij to maintain the magnitude relationship of the key value as a character string, and add the item identification Cj and data management number Rx is added, and a value "1" corresponding to "addition" is added as an addition/deletion identification to create index update information kiJ, and index update information output processing S8
Transfer control to

FIGS. 5(b) and 5(e) show the index item f4. of the additional data Di in FIG. Update information di4. corresponding to f5. di5 key value 111111. An example of creating index update information k14 and ki5 as 100111110°99999977777 by adding the maximum value 99999999.99999999999 that can be taken based on the item definition to -22222 is shown.

Note that character type key value conversion processing S5. Unpacked format 10
Decimal key value conversion process 36. The key value of the index update information ktj created in packed format decimal key value conversion process S7 is
Its length is adjusted to the maximum of all index items, padded with spaces on the right.

(8) Index update information output processing S8 The index update information kiJ is output to the work file 3 for sorting.

(9) Item end determination processing S9 Compare j with the last index item number in the additional data, and if j is smaller than that, there are unprocessed index items remaining, so add 1 to j and extract index update information Control is passed to process S3, and if j is a value greater than that value, all items are considered to have ended and control is passed to update data end determination process S10.

(10) Update data end determination process 5 Compare IO1 with the number of update data (in this case, the number of additional data) included in update data file 4, and if I is smaller than that, there is unprocessed additional data remaining. 1 is added to , and control is passed to the updated data importing process Sl. If i is a value greater than that, control is passed to the batch sorting process Sll as the end of all additional data.

(11) - Batch sort processing Sll Add/add all the index update information output to the work file 3 for sorting, using the item identification as the first key, the key value as the second key, and the data management number as the third key. Sort all at once using the deletion identification as the fourth key.

(12) Index update information S12 From the work file 3, sequentially input the index update information arranged in key order for each index item by the -batch sort processing Sll, and execute key insertion into the corresponding index section of the mask file 5. This completes the series of updates.

The operation of this embodiment has been explained above using data addition processing as an example. However, in the case of data deletion processing, index update information kj to which identification value 2 indicating "deletion" is added as addition/deletion identification is Created. In addition, in the case of data change processing, index update information kij is created regarding the old data to which identification value 2 indicating "deletion" is added as the addition/deletion identification, and index update information kij is created that indicates "addition" as the addition/deletion identification. Index update information kiJ regarding new data to which index update information klj to which identification value l has been added is created.

〔Effect of the invention〕

As explained above, in the index update method of the present invention, update information including key values extracted for multiple index items included in each update data is batched in an order suitable for updating the index for each index section. The index update information is converted into index update information consisting of character strings in a sortable format and stored in an intermediate file, and the index update information stored in the intermediate file is sorted all at once, making it suitable for updating each index section. Since index update information is obtained in order, the intermediate files that were conventionally required for each index item can be reduced to one, and the overhead when outputting index update information to intermediate files can be reduced. It becomes possible.

Therefore, the index update processing time can be shortened.

[Brief explanation of the drawing]

FIG. 1 is a flow chart of processing related to index update in an embodiment of the present invention, FIG. 2 is a block diagram of main parts showing an example of an information processing system to which the index update method of the present invention is applied, and FIG. 3 is a mask file 5 FIG. 4 is a diagram showing an example of the format of additional data, and FIG. 5 is a diagram showing an example of the conversion format of index update information. In the figure, 1... Processing device 2... Randomly accessible storage device 3... Sorting work file (intermediate file) 4... Update data file 5... Mask file 6... Data Part 7-1~? -n...Index section Processing related to index update in the embodiment of the present invention FIG. 1 (
1)

Claims (1)

[Claims] A file having a data section and a plurality of index sections storing index data including key values of index items of data stored in the data section and location information of the corresponding data for each index item is randomly generated. In an index update method in an information processing system that stores the data in an accessible storage device and updates the data section based on a series of update data, and then updates the index section related to the current series of updates at once. ,
Update information including key values extracted for multiple index items included in each update data is converted into index update information consisting of strings in a format that can be sorted all at once in an order suitable for updating the index for each index section. comprising a converting means for converting, and an intermediate file for storing the index update information generated by the converting means, and after sorting the index update information stored in the intermediate file in the above order at once, An index update method characterized in that the index section is updated according to contents.