JPH05334354A - Method for reading-in index organization file - Google Patents

Abstract

PURPOSE:To access a file at a high speed by searching for the key right before a current key by a binary search method and accessing keys in the decreasing order. CONSTITUTION:When the key right before the current key is searched for in a constitution example of a file consisting of a key part 12 and a data part 13, the binary search method is applied to narrow down the range of the search to 1/2, 1/4, 1/8.... Namely, when the current key is larger than the value of HIGH-KEY, the target key right before it is positioned before an intermediate point, so W-KEY (intermediate point) is transferred to HIGH-KEY (upper limit). When the current key is smaller than HIGH-KEY, the target key right before the current key is present between the intermediate point and HIGH-KEY, so the lower limit of the binary search is raised. At this time, the frequency of search for narrowing down the range is made less than that of a normal binary search, so the current key is transferred to not the intermediate point, but LOW-KEY.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a file reading method in an information processing system, and more particularly to a reading method for using an index organization file.

[0002]

2. Description of the Related Art The conventional reading method for indexed files is as follows. Directly read the record of the specified key 1. The method of reading the records in the order of the keys (ascending order) is common, and JIS COBCL also provides instructions. However, from the value of a key to the reverse order (descending order)
When reading a record into, the value of the key is read as the value of the key to be read next by subtracting 1 from the value of the current key.
While subtracting each one, a method of repeatedly reading until a record was found was used.

[0003]

In the above method, when the difference between the current key and the immediately preceding key is large, the number of times of reading is increased until the key is found, and the processing time becomes long.

[0004]

According to a first aspect of the present invention, in a method of reading an index organization file in a system that uses an index organization file, the key immediately before the current key is searched for by the binary search method, and the key is searched. It is characterized by accessing in descending order.

[0005]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described with reference to the drawings.

FIG. 1 is a flow chart showing an embodiment of the present invention.

The current key existing in the index organization file (hereinafter referred to as a file) to be searched is given. The procedure for finding the key immediately before the current key will be described below. The search method is an application of the binary search method, and is a method of narrowing the search range to ½, ¼, ⅛ ...

First, the current key is stored in the work area W-current and HIGH-KEY (upper limit of binary search) (step 1). 0 is transferred to the current key (step 2). Position the file at the current key and read the record. The key obtained here is the minimum key of this file (step 3). Compare the W-current stored in step 1 with the minimum key of the file obtained in step 3, and if W-current is equal to the minimum key, the current key is the minimum key of the file, so Since the previous key does not exist, the search ends (step 4). Only if not equal to the next step. LOW-K the current key read in step 3
Transfer to EY (lower limit of binary search) (step 5). L
The intermediate point between OW-KEY and HIGH-KEY is calculated and obtained, and the calculation result is transferred to the current key and W-KEY (step 6). LOW-KEY (lower limit) and W-KEY (midpoint) are compared, and when they are equal, this midpoint W-KEY
Is the last key (step 7). If not equal, proceed to the next step 8.

The reason why the intermediate point W-KEY is the key immediately before the purpose will be described below. W-KEY is an n-digit integer item. When dividing, the numbers after the decimal point shall be truncated. Binary search is LOW-KEY and HIGH-K
This is a search method that narrows the space between EY, but the difference between LOW and KEY immediately before the target key is found is 1. Assuming that m is a target key, (LOW-KEY tens HIGH-KEY) / 2 = W-KEY = (m tens (m tens 1)) / 2 = m tens 0.5. Here, 0.5 is rounded down, so W-KE
M is substituted for Y. Therefore, LOW-KEY =
When W-KEY, it is the discovery of the target key.

When LOW-KEY and W-KEY are not equal, the file is positioned at the intermediate point W-KEY, and the record of the closest (largest) key above the intermediate point value is read (step 8). Since the read key is stored in the current key, it is compared with HIGH-KEY (step 9). When the current key is higher than the value of HIGH-KEY, the key immediately before the target is located before the intermediate point, so W-KEY (intermediate point) is transferred to HIGH-KEY (upper limit) (step 10). ). Currently the key is HIGH
If it is smaller than – KEY, the midpoint and HIGH – KEY
Since the key just before the target exists between and, the lower limit of the binary search is raised. At this time, the current key is moved to LOW-KEY instead of the intermediate point in order to narrow the range and reduce the number of searches compared to the normal binary search (step 11). Steps 6 to 11 are repeated until the previous key is found in step 7.

FIG. 2 is a diagram showing an example of the structure of a file consisting of the key part 12 and the data part 13.

Next, a process of searching the previous key "075" from the current key "100" in the file shown in FIG. 2 will be described. 1. KEY = "0" to get the smallest key on the file
Position with "00" and READ. KEY = "00"
"1" is read, but "100" is not equal, so "
It can be seen that 100 ″ is not the smallest key.

LOW-KEY = "001" HIG
H-KEY = “100” 2. (LOW-KEY 10 HIGH-KEY) = /
Since 2 = 50.5 → 50 is the intermediate point, KEY = ″
When positioned at 050 "and read," 055 "is read.

Since "055" <"100", "055" is transferred to LOW-KEY.

LOW-KEY = "055" HIG
H-KEY = “100” 3. When the intermediate point is calculated, it becomes 77, so KEY = "0.
If you position and read with "77", "100" will be read. Since the current key and HIGH-KEY are equal, H
Transfer "077" to IGH-KEY.

LOW-KEY = "055" HIG
H-KEY = “077” 4. When the intermediate point is calculated, it becomes 66, so KEY = "0.
"075" is read when positioned at 66 "and read." 075 "<" 077 ", so" 075 "is transferred to LOW-KEY." 075 "HIGH-KEY =" to LOW-KEY.
077 ″ 5. The intermediate point is calculated to be 76, so KEY = ″ 0
When the position is read at 76 "," 100 "is read. Since" 100 "≥" 077 ", the intermediate point" 076 "is transferred to HIGH-KEY.

LOW-KEY = "075" HIGH
-KEY = "076" 6. The intermediate point is calculated to be 75, so LOW-KE
It becomes equal to Y and the repetition ends. The intermediate point "075" at this time is the immediately preceding key.

[0018]

As described above, according to the present invention, in the example of FIG. 2, the key immediately before can be found by READ 5 times.
If you read by subtracting the key one by one, it will be 25 times RE
You have to do AD. Therefore, when the files are read one after another in descending order, there is an advantage that the file can be accessed at a higher speed.

[Brief description of drawings]

FIG. 1 is a flowchart showing an embodiment of the present invention.

FIG. 2 is a diagram showing an example of the structure of a file in this embodiment.

[Explanation of symbols]

 12 key part 13 data part

Claims (1)

[Claims] 1. A method of reading an index organization file in a system using an index organization file, characterized in that a key immediately before a current key is searched based on a binary search method to access the keys in descending order. How to read the organization file.