JPS63148326A - Data coupling device - Google Patents

Abstract

PURPOSE:To reduce the amount of hardware of a device by determining data to be next compared on the basis of a movement destination determining table of memory read address and duplicate data group start address in accordance with current and preceding comparison results between two data sets. CONSTITUTION:One word of data of each of data sets A and B is inputted from each of input terminals 201 and 200 and is written in each of memories 203 and 202. One word of data is read out from each of memories 203 and 202 to each of memories 205 and 204 and each of memories 208 and 207 simultaneously, and read-out words are compared by a comparator 206 to discriminate whether one of these words is larger than, smaller than, or equal to the other, and next data is inputted to the data set where the value of a common item is smaller, and it is written in the memory of this data set side. Meanwhile, if data whose common items coincide with each other are found as the comparison result, one word of data stored in each of memories 208 and 207 is read out and is outputted from an output terminal 213.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention is a data combination device, that is, in data processing that manipulates data sets, it extracts data with matching values of the common items from two data sets having common items. This invention relates to a device that finds and outputs pairs.

[Conventional technology]

For example, when finding data with matching common item values from two data sets, one data set (data set B) has multiple data items (duplicate data group) with the same common item values, and the other data set (Data set A) may have one or more pieces of data that have the same common item value.

Assuming that each data set is sorted by the value of a common item, the comparison to find matching data in this case is as shown in FIG. Here, 1 is data set A, 2 is data set B, 11 to 13 are data of data set A, and 21
~24 are data of Data Set B, 31 are common items of Data Set A, 32 are items that are not common to Data Set A, 41 are common items of Data Set B, and 42 are common items of Data Set B. Items that are not common in data, 5
1 to 58 indicate data comparison steps; 51,
52. It is assumed that the comparison is performed in the order of... Further, each data set is assumed to be arranged by common items 31 and 41, and data 22 and 23 are a duplicate data group.

In this comparison, data 2
2.23' is known to be a duplicate data set, and using this information the comparison step 57 can be omitted. Generally, there are duplicate data groups in a data set,
It is important to reduce this comparison step.

Conventionally, in order to reduce the number of comparison steps, a method has been used in which a mark is attached to duplicate data in advance and the mark is used to prevent unnecessary comparisons. For the purpose of finding the above-mentioned duplicate data, for example, the relational database machine Delta J (Nikkei Electronics, 1985.9.23tPP-235-280) equipped with an engine dedicated to r-relational algebraic operations by Tsunoda et al.
r A Hardwa by S, Kamiya et al.
rePipeline Algorithm for
Relational Data-base
Processing and Its Impl
mentationUsing a Dedica
ted HardwareJ (I COTTec
hnical Report: T R-091N
A dedicated circuit such as that described in J. Ove+mber, 1984) was used.

In FIG. 7, which shows an example of the configuration of a conventionally proposed device, 71 is a duplication determination section, 72 is a coupling section, 73, 74
are both data 1 of data set A or data set B
Memories 75 and 76 each store one word of the data stored in the memories 73 and 74 (for example,
77 is the register 7 above.
5. A comparator that compares the magnitude of the contents of 76; 78 a selector that outputs the contents of either register 75 or 76 to the coupling section 72; This is an address control unit that controls .74 write/read addresses.

Further, 80.81 is a memory for storing data belonging to the data sets A and B, respectively, and 82.83 is a memory 80, respectively.
．． A register 81 stores one word of the stored data, 84 a comparator that compares the magnitude of the contents of the registers 82 and 83, and 85 and 86 each store one data belonging to the data sets A and B. Memory 87 is the memory 85.
A selector 88 outputs the contents of one of the memories go and a based on the comparison result of the comparator 84.
This is an address control unit that controls the writing/successive address of t.

Among the basic processing procedures of the duplication determination section 71 configured as described above, the duplication flag adding processing procedure is shown in FIG. Here, 100 to 102 are data of data set A, 104
is a common item between data sets A and B, 105 is data set A
106 is an item that is not common between datasets A and B, 110 to 113 are data of dataset B, 114 is a common item between datasets A and B, 115 is a duplicate flag storage area of dataset B , 116 are items that are not common between data sets A and B, and 120 to 122 are items 100 to 122, respectively.
As a result of assigning a duplicate flag to 102, 127 is the address where data set A is written in the memory 80, 130 to 13
3 is the result of adding duplicate flags to 110 to 113, respectively.
137 is the address where the data set B is written in the memory 81, and 140.141 is the data of the combination result. Here, 100 to 102 are already aligned for 104, 110 to
It is assumed that 113 has already been aligned with respect to 114. The processing procedure is shown below.

■The duplication determination unit 71 determines that data 100 of data set A
Regarding ~102, when the value of the common item of each data is equal to the value of the data next to the data, a process of storing "1" in the duplication flag area of the former data, that is, a process of assigning a 1 duplication flag is performed. , stored in the memory 80 of the coupling unit 72. 120 to 122 in FIG. 8 show the stored state.

■The data 110 of data set B is determined by the duplication determination unit 71.
For ~113, the same process as ■ is performed, and the joining part 7
The results are stored in the memory 81 of No. 2. At this time, when the data is sent to the 1-data combining section 72, the following process (2) is started.

(2) In the combining unit 72, the data stored in the memory 80 and the memory 81 are compared for common items from the beginning, and data having the same value are successively output.

The order of comparison is shown in FIG. Here, 151 to 155 are data comparison steps. When comparing the duplicate data groups of data set B, the duplicate flag is used to reduce the number of comparison steps as follows. That is, at the end of the comparison in step 154.

Since the comparison results in this step were equal and "1" was not set in the overlap flag area for both data 121 on the data set A side and data 132 on the data set B side, the next comparison will be made for each data set. Next data 122 and 133
I judge that.

[Problem that the invention seeks to solve]

In the above-mentioned conventional technology, in order to reduce the number of comparisons in duplicate data groups, a method is used in which a mark (flag) is attached to the duplicate data groups in advance, but a duplicate determination section 71tt is provided to detect duplicate data. Therefore, there was a problem in that the amount of hardware for the device increased.

The present invention has been made in view of the above circumstances, and its purpose is to solve the above-mentioned problems in conventional data combining devices, and to solve the problems of the device's hardware without increasing the processing time for comparing duplicate data groups. An object of the present invention is to provide a data combination device that can significantly reduce the amount of software.

[Means for solving problems]

The above-mentioned object of the present invention is that when two data sets are each composed of a plurality of items having a part in common, and each data set is already sorted by the common items, a common item is extracted from these data sets. A data combination device that finds data pairs with matching item values, connects them, and outputs them.
two memories for storing respective data sets, two storage means for storing data read from each of the memories or a common item of data, and a size of one word of the common item of data in the two storage means. A comparator that compares / etc. / Small;
a selector connecting either of the two storage means and a data output terminal of the device, and controlling each of the means from the current comparison result by the comparator and the comparison result immediately before the comparison;
This is achieved by a data combining device characterized in that it is provided with a control means for determining the data to be compared next.

[Effect]

In the present invention, data to be compared next is determined from the current comparison result and the previous comparison result of two data sets, based on a movement destination determination table for memory read address 9 duplicate data group first address, which will be described later. It was decided to do so.

Therefore, there is no need to determine whether there is duplicate data, no circuit for determining duplicates is required, and there is an advantage that the amount of hardware in the device can be reduced. In addition, in the case of the apparatus of the present invention, the number of comparisons is greater than that of the conventional apparatus, but since the data is output sequentially word by word, including items that are not common, the comparison time is overlapped by this output time, and the comparison process is The time will be the same as before.

〔Example〕

EMBODIMENT OF THE INVENTION Below, one embodiment of the present invention will be described in detail based on the drawings.

FIG. 1 is a block diagram of a data combining device that is an embodiment of the present invention. In the figure, 200 and 201 are input terminals of the data sets B and A, respectively, and 202 and 203 are memories that store data of the data sets B and A, respectively.

204 and 205 are registers each storing one word of data to be read from the memory 202 and 203, 206 is a comparator for determining whether the data is large or small, and 207 and 208 are 2 data size registers for storing the data being compared. memory, 210 is a selector, 2
13 is an output terminal, and 214 is a control circuit.

The main pointers in the memory 202.203.207.208 are as follows.

(a) Write pointer (W pointer) [for A, B]:
Each indicates a written data location within memory 203 and 202.

(b) Read pointer (R pointer) [Indicates the position of the next data to be read in the memories 203 and 202 for A and B, respectively.

(C) Output pointer (0 pointer) [for A, B]: Indicates the position of output data in the memory 208 and 207, respectively.

(d) Current duplicate data group pointer: Points to the position in the memory 202 where the first data of the duplicate data group to which the data currently being read in data set B belongs is stored.

(e) Next duplicate data group pointer: Indicates the position in the memory 202 where the first data of the next duplicate data group to be read out of the duplicate data group to which the data currently being read in data set B belongs is stored.

The data combining device shown in this embodiment has two clocks: a write clock (W clock) and a read clock (R clock).
Process with clock. The basic processing using the above W and R clocks is as follows.

[One word of the data of data sets A and B is input from the W clock input terminals 201 and 200, respectively, and written to the memories 203 and 202. At the same time, if there is data with matching common items, it is stored in the memories 208 and 207. any one of the data
The word is read out and output from the output terminal 213.

Here, writing and output are performed based on the following basic rules.

〔writing〕

(a) As a result of the comparison with the comparator, input the next data under the conditions of (b) to the data set with the smaller value of the common item (even if data set B is equal), and Write to collective memory.

(b) If the input of the previous data has not been completed, wait to input the next data until the input/writing of that data is completed.

(c) Increase the ◯ pointer by "1" each time one word is written.

〔output〕

(a) As a result of the comparison, the data in the memories 207 and 208 is output word by word from the next process when it is found that the values of the common items are equal.

(b) Increase the ○ pointer by "1" each time one word is output.

[R Clock] One word of data is read from the memories 203 and 202 into the registers 205 and 204 and the memories 208 and 207 at the same time, and the comparator 206 determines whether it is larger/smaller/equal.

Here, reading is performed based on the following basic rules.

(Continued) (a) Reading is performed simultaneously from two memories (202, 203).

(b) In the case of a common item word, if the word to be compared is present in two memories under the condition (c), that word is read out. For terms of items other than common items, whether the values of the common items are equal or
Alternatively, if a 1 judgment has not yet been made, it is read out.

(Q) If a word different from the word you are currently trying to read is stored in the storage location of the memory 207 or 208,
Wait for reading until the word is output.

(d) Each time one word is read, the R pointer is incremented by "1", and at the time when the comparison judgment result is obtained or when one data readout is completed, according to the contents of the table shown in FIG.
The R pointer, the current duplicate data group pointer, and the next duplicate data group pointer are moved. In FIG. 2, the data to be compared next is determined using the current comparison result and the previous comparison result.

FIG. 3 is a diagram illustrating the comparison step in the data combination device of this embodiment, and 301 to 307 indicate data comparison steps.

FIG. 4 is a diagram showing the flow of processing, and 400 and 401 are one word of data of data sets B and A, respectively.

Data set A is (2,b), (3,c), (4,a),
Data set B is Dt t)t(at 5)t(at v
) and (5, u) (see Figure 3). However, the first item in parentheses is a common item between the two data sets, and the second item is a non-common item. It is also assumed that data input is performed independently for each set from a sequential file such as a magnetic tape.

FIG. 4 shows the above W for each of times [1] to [12].
This shows the write, output, and read processing using the clock and R clock in sequence, and the symbols 200 to 20
8.212.213 indicates the same components as previously shown in FIG.

As can be seen from a comparison between FIG. 9 and FIG. 3, the number of comparisons is greater in the processing performed by the apparatus of this embodiment than in the case of the conventional apparatus. This is because, in the case of this embodiment, data is sequentially output word by word, including items that are not common. Then, the comparison time is superimposed on this output time, and the comparison processing time is the same as before.

Figures 5(a) and 5(b) are diagrams for explaining this.
3 shows the processing performed by the conventional device (a) and the device (b) of the present embodiment in each case. The data used here is that data set A is (3, c), (4, a
), and data set B is (3,s).

(5, u). Furthermore, in order to compare only the processing times, the processing method of the conventional device is matched to the processing method of the device of this embodiment.

In FIG. 5, the processing of the device of this embodiment and the conventional device is as follows:
Both end at time [5]. If there are two or more pieces of data with matching common item values in each data set, the output time becomes even longer, and the comparison time is superimposed on this.

〔Effect of the invention〕

As described above, according to the present invention, when each of two data sets is composed of a plurality of items having a part in common, and each data set is already sorted by the common items, these A data combination device that finds, connects, and outputs data pairs with matching common item values from data sets, and includes two memories that store each data set, and data read from each of the memories or a common item of data. two storage means for storing, a comparator for comparing the magnitude/equal/smallness of one word of a common item of data in the two storage means, and one of the two storage means and a data output terminal of the device. and a control means for controlling each of the means from the current comparison result by the comparator and the comparison result immediately before the comparison and determining the data to be compared next. This has the remarkable effect of realizing a data combining device that can significantly reduce the amount of hardware in the device without increasing the number of times data groups are compared.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a data linking device showing an embodiment of the present invention, FIG. 2 is a diagram explaining how each pointer moves in the embodiment device, and FIG. 3 is a diagram showing a comparison step in the embodiment. , FIG. 4 is a diagram showing the flow of processing, FIG. 5 is a diagram comparing the flow of processing between the conventional device and the embodiment, and FIG. 6 is a diagram illustrating the step of comparing data sets with duplicate data groups.
FIG. 7 is a diagram showing the configuration of a conventional data combining device, FIG. 8 is a diagram showing an overview of the processing flow of the conventional device, and FIG. 9 is a diagram showing comparison steps in the conventional device. 1: Data set A, 2: Data set B, 11 to 13: Data of data set A, 21 to 24: Data of data set B, 31: Common items of data of data set A, 32: Data of data set A items that are not common. 41: Common items in the data of data set B, 42: Items that are not common in the data in data set B, 200, 201
: Input terminals of data sets B and A, 202, 203 : Memory for storing data of data sets A and B, 204, 2
05: Register that stores one word of data, 207.2
08: Memory for storing data under comparison, 212: Selector. 213: Output terminal, 214: Control circuit, 301 to 307
: Data comparison step, 400, 401 : One word of data of data sets A and B. Go to Figure 3
−≧ F
Yen≧ C - 6th
Figure 8

Claims (1)

[Claims] (1) When two data sets each consist of a plurality of items that have a part in common, and each data set has been sorted by the common item, the matching of the values of the common items from these data sets A data combination device for finding, concatenating, and outputting data pairs, comprising two memories for storing respective data sets, and two storage means for storing data read from each of the memories or common items of the data; 1 of the common items of data in the two storage means
a comparator that compares the magnitude/equal/smallness of words; a selector that connects either of the two storage means and the data output terminal of the device; A data combining device characterized by comprising a control means for controlling each of the means based on the comparison result and determining data to be compared next.