KR101429046B1 - Method for searching, inputting, deleting and garbage collecting of data in database having key-value structure - Google Patents

Abstract

Input, deletion, and garbage collection methods that can prevent the occurrence of errors due to the asynchronism of the forward index and the backward index while simultaneously improving the data retrieval speed by storing the forward index and the reverse index simultaneously do. A method for retrieving data in a database having a key-value structure according to the present invention is a method for retrieving data from a database having a key-value structure, comprising the steps of: storing inverse index data having a combination of a field name and a field value of a data as a key and a combination of an identification value and a time stamp of the data as a value item; A method for searching data in a database having a key-value structure, storing forward index data having as a key an identification value of the data and having a combination of a field name, a time stamp, and a field value of the data as a value item, (a) receiving a field name and a field value of data to be searched, (b) searching the database for inverse index data having a combination of a field name and a field value of the search target data as a key, (c) Forward index data having an identification value included in a value item of the inverse index data as a key is stored in the database And (d) returning the searched forward index data as a search result, wherein a value item having the latest time stamp among the value items of the searched forward index data having the same field name as the field name of the search object data And returning the searched forward index data as a search result only when the field value has the same field value as the field value of the search object data.

Description

METHOD FOR SEARCHING, INPUTTING, AND DELETING AND GARBAGE COLLECTING OF DATA IN DATABASE HAVING KEY-VALUE STRUCTURE In a database having a key-value structure,

The present invention relates to a method for searching, inputting, deleting and garbage collecting data in a database having a key-value structure, more specifically, searching, inputting and deleting data simultaneously and concurrently by a plurality of commands Input, deletion, and garbage collection of data without conflicts of operations on concurrent commands in the database.

As the era of big data comes, massive data processing is becoming an issue. Since existing relational databases (RDBs) are not easy to expand horizontally and require schemes that are dependent on relationships, they are difficult to apply to storage and management of large and diverse types of data . In contrast, the NoSQL database is becoming a mainstream database in the era of big data because it is easy to expand and easily store and manage unstructured data.

Currently, various kinds of NoSQL databases such as Cassandra, Mongo DB, Riak, and H-Base have been proposed, and most of these NoSQL databases have a key-value structure.

Fig. 1 (a) shows an example of data having such a key-value structure. In the example of FIG. 1 (a), employee numbers '101' and '102' are keys, and a value dependent on a colon (:) for each key is a value. That is, the contents of braces ({}) are the values for each key. Thus, in a key-value structure, one value is dependent on one key. Also, one value may have one or more value items. In the example of Fig. 1 (a), "Name: Hong Kil Dong" and "Age: 26" are each a value item. At this time, each value item has a field name of 'Name' or 'Age' and a field value of 'Hong Kil Dong' or '26' in a key-value structure.

Since retrieval of data in a database having such a key-value structure depends on a key, it is difficult to retrieve data depending on the value. For example, in the data shown in Fig. 1 (a), since the employee number is a key, it is easy to retrieve the data of the employee having the employee number 101, but the employee whose name is Hong Kil-dong or the employee whose age is 26 It becomes difficult to search for a value dependent value.

1 (a) is set as the key, and the key of the data of Fig. 1 (a) is set as the value item, that is, the key-value structure is reversed to create the inverted index There is an attempt to facilitate searching for a value item.

1 (a) is referred to as a forward index, FIG. 1 (b) is a reverse index of the forward index in FIG. 1 (a). Inverse index has a value item of the forward index as a key, so it is possible to search depending on name or age. For example, an employee whose age is 26 can easily search from the inverted index of FIG. 1 (b).

As such, if the database has an inverse index, there is an advantage that the search speed is significantly increased. However, since data must be stored in both the forward index and the backward index, there is a disadvantage in terms of the storage capacity. In particular, there is a problem that the forward index and the backward index must be accurately synchronized with each other.

However, since addition, update, and deletion of a forward index and a reverse index are not strictly concurrently performed on one piece of data, after either the forward index or the backward index for one piece of data is changed and before the other is changed When the addition, update, deletion, or retrieval of the one data is performed, an error due to asynchronism between the forward index and the reverse index may occur.

In particular, in the case of big data managing large amounts of data, there is a high possibility that errors due to asynchronization of the forward index and the backward index are likely to occur at the same time because the addition, update, deletion and retrieval of data are likely to occur simultaneously.

Disclosure of Invention Technical Problem [8] The present invention has been made to solve the above-mentioned problems, and it is an object of the present invention to provide an apparatus and a method for simultaneously storing a forward index and an inverse index, , Data retrieval, input, deletion, and garbage collection methods.

According to another aspect of the present invention, there is provided a method of retrieving data from a database having a key-value structure, comprising the steps of: using a combination of a field name and a field value of a data as a key, And the forward index data having a combination of a field name, a time stamp, and a field value of the data as a key, and forward index data having an identification value of the data as a key and storing the forward index data in a database having a key- (A) inputting a field name and a field value of data to be searched, (b) searching the database for inverse index data having a combination of a field name and a field value of the search target data as a key (C) a forward index having a key as an identification value included in the value item of the retrieved inverse index data (D) returning the retrieved forward index data as a retrieval result, wherein a latest timestamp among the value items of the retrieved forward index data having the same field name as the retrieval object data field name And returning the retrieved forward index data as a retrieval result only when the value item having the field value having the same field value as the field value of the retrieval object data.

According to another aspect of the present invention, there is provided a method of inputting data in a database having a key-value structure, comprising the steps of: obtaining a combination of an identification value and a time stamp of the data, Value structure having a key-value structure in which, as a value item, inverse index data possessed as a value item and a combination of a field name, a time stamp, and a field value of the data with the identification value of the data as a key are stored. (A) inputting an identification value, a field name, and a field value of input data, (b) using a combination of a field name and a field value of the input data as a key, and identifying Value and a time stamp as a value item to the database, and (c) and inputting the forward index data having the identification value of the input data as a key and having a combination of a field name, a time stamp, and a field value of the input data as a value item, when the step (b) is successful.

According to another aspect of the present invention, there is provided a method of deleting data in a database having a key-value structure, the method comprising: storing a combination of an identification value and a time stamp of the data, Value structure having a key-value structure in which, as a value item, inverse index data possessed as a value item and a combination of a field name, a time stamp, and a field value of the data with the identification value of the data as a key are stored. The method comprising the steps of: (a) receiving an identification value and a field name of data to be deleted; and (b) inputting, to the forward index data having the identification value of the data to be deleted as a key, A combination of a field name, a time stamp, and a field value indicating that the data is to be deleted is added as a value item .

According to another aspect of the present invention, there is provided a method for garbage collection of data in a database having a key-value structure, comprising the steps of: Value structure having a value field and a combination of a field name, a time stamp, and a field value of the data with the identification value of the data as a key, and a forward key index A method for garbage collection of data in a database, the method comprising the steps of: (a) if the forward index data in the database has two or more value items having the same field name, a garbage value item having no latest timestamp among the two or more value items Of a field value and a combination of field names as keys (B) deleting a value item having a key of the forward index data as an identification value in the retrieved inverse index data; and (c) deleting the value item of the forward index data from the forward index data .

According to the present invention, it is possible to improve the retrieval speed of data by simultaneously storing the forward index and the reverse index, and at the same time to prevent the occurrence of errors due to the asynchronization of the forward index and the reverse index at the time of retrieving, It is effective.

1 shows an example of data having a conventional key-value structure.
2 shows an example of forward index data according to an embodiment of the present invention.
FIG. 3 shows an example of inverse index data according to an embodiment of the present invention.
4 is a block diagram showing a schematic configuration of a system in which a method of inputting, deleting, searching, and garbage collecting data in a database having a key-value structure according to an embodiment of the present invention is performed.
5 is a flowchart illustrating a method of inputting data in a database having a key-value structure according to an embodiment of the present invention.
6 to 11 show an example of a data input process according to an embodiment of the present invention.
12 is a flowchart illustrating a method of deleting data in a database having a key-value structure according to an embodiment of the present invention.
13 shows an example of a process of deleting data according to an embodiment of the present invention.
FIG. 14 is a flowchart illustrating a method for searching data in a database having a key-value structure according to an embodiment of the present invention.
15A to 17 illustrate an example of a data retrieval process according to an embodiment of the present invention.
18 is a flowchart illustrating a method of garbage collection of data in a database having a key-value structure according to an embodiment of the present invention
FIG. 19 shows an example of a garbage collection process according to an embodiment of the present invention.

BRIEF DESCRIPTION OF THE DRAWINGS The advantages and features of the present invention, and the manner of achieving them, will be apparent from and elucidated with reference to the embodiments described hereinafter in conjunction with the accompanying drawings. The present invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art. To fully disclose the scope of the invention to those skilled in the art, and the invention is only defined by the scope of the claims. Like reference numerals refer to like elements throughout the specification.

Hereinafter, a method for inputting, deleting, searching, and garbage-collecting data in a database having a key-value structure according to an embodiment of the present invention will be described with reference to the drawings.

<Database Configuration>

The database used in the present invention is a database having a key-value structure. Preferably, it can be a NoSQL database that is easy to manage and scalable with a vast amount of data. Each piece of data in a database of key-value structures can have a hierarchical key-value structure. That is, one value corresponding to one key may be composed of one or more value items having a key-value structure.

2 (a) shows an exemplary schema of a forward index in the database of the present invention, and Fig. 2 (b) shows an exemplary forward index data having a schema of Fig. 2 (a).

Referring to FIG. 2 (a), a document used as a unit of data in the database of the present invention has one document identification value as a key and one or more sets of value items as one value. Each value item has a key-value structure. In the example shown in Fig. 2 (a), the combination of the field name and the time stamp becomes the key of each value item, and the field value becomes the value of each value item. Here, the time stamp is a value indicating when data is input, and more specifically, it is a value indicating when a value item, that is, a field name and a field value, is input.

Since the key-value database is key-dependent, there can not be more than one data having the same key in the same layer. For example, in the value layer of '<Document ID A>' in FIG. 2, there can not be two or more value items having '<Field Name A>' as a key. However, there is a possibility that data having the same key may occur in the big data in which data input occurs simultaneously.

For example, as in the example shown in FIG. 2 (b), the field data of "name: doody" is input to the document of the employee number "101" at the time of "ts1" : &Lt; / RTI &gt; &lt; RTI ID = 0.0 &gt; If the key of each value item is made up of the field name only, there can not be two or more value items having a key of 'name' for the document of the employee number '101'. However, in the example of FIG. 2 (b), since each value item has a combination of a field name and a time stamp, the field data of 'Name: Doodle' and the name of ' Field data can be present at the same time.

3 (a) shows an exemplary schema of an inverted index obtained by inverting the forward index in FIG. 2 (a), and FIG. 3 (b) shows an inverse index data for the forward index data in FIG. 2 .

In the inverse index, the combination of the field name and the field value, which became the value items of the forward index, is the key, and the document identification value that became the key of the forward index becomes the value item. More specifically, the value item has a combination of the document identification value and the time stamp as the key of the value item.

The forward index data illustrated in FIG. 2B may be stored in the forward index database, the reverse index data illustrated in FIG. 3B may be stored separately from the reverse index database, and the forward index data and the reverse index data may be stored in the same It may also be stored in a database.

<System Configuration>

4 is a block diagram showing a schematic configuration of a system in which a method of inputting, deleting, searching, and garbage collecting data in a database having a key-value structure according to an embodiment of the present invention is performed.

The system 100 includes a data input unit 110, a data search unit 120, a garbage collection unit 130, and a database 140. The system 100 is connected to the user terminal 200 through a network and can receive input data, deletion target data, or search target data. In addition, the system 100 may be connected to any terminal or server 300 on the network and receive input data by a crawler or the like.

A method for inputting and deleting data in a database having a key-value structure according to an embodiment of the present invention is basically performed by an additional operation of data. Therefore, data input and deletion need not be performed by separate modules or engines, respectively, and all of them can be performed by the data input unit 110. [

The data input unit 110, the data retrieval unit 120, the garbage collection unit 130, and the database may be configured as one physical device or a combination of a plurality of devices. In particular, the database 140 may comprise a distributed database in which a plurality of devices are connected via a network.

Further, a method for inputting, deleting, searching, and garbage collecting data in a database having a key-value structure according to an embodiment of the present invention can be performed by a computer program stored in a recording medium such as a hard disk of the system.

<Method of adding data>

FIG. 5 is a flowchart illustrating a method of inputting data in a database having a key-value structure according to an embodiment of the present invention, and FIG. 6 illustrates an example of a data input process according to an embodiment of the present invention. Fig. 6 schematically shows the forward index data and the backward index data illustrated in Figs. 2 (b) and 3 (b).

First, the data input unit receives the identification value, the field name, and the field value of the input data (S110). The identification value, the field name, and the field value of the input data may be input by the user, or may be collected and input from any terminal or server connected to the network by a crawler or the like. In the example of Fig. 6, the database storing the forward index data and the backward index data illustrated in Fig. 2 (b) and Fig. 3 (b) Input data having a 'donor' (field value) and an age (field name) '28' (field value) are input.

At this time, it is assumed that the time stamp indicating the time of inputting the input data is '<ts 4>'. The time stamp may be generated according to the date and time when the input of the user or the crawler reaches the data input unit or may be generated according to the date and time when the input is transmitted from the user terminal or the crawler to the data input unit It is possible.

The data input unit adds the inverse index data having the combination of the field name and field value of the input data as a key and the combination of the identification value and the time stamp of the input data as the value items to the database (S120). For example, in FIG. 6, with the key having the name = donor as the key, the inverse index data having the combination of the employee number '103' and the time stamp '<ts4> Inverse index data having a combination of the employee number '103' and the time stamp '<ts4>' as a value is added to the database (S120).

If the addition of the inverse index data is successful, the data input unit adds the forward index data having the identification value of the input data as a key and having a combination of the field name, the time stamp, and the field value of the input data as the value items to the database (S130) . For example, in FIG. 6A, a combination of 'Name', '<ts 4>' and 'Donner' and 'Age', '<ts 4>' and '28' Are added to the database (S130).

If the addition of the inverse index data fails, the data input unit returns the data addition failure information without adding the forward index data to the database (S140). In addition, even if the addition of the forward index data fails, the data addition failure information is returned (S140). The data input unit returns data addition success information only when both the addition of the inverse index data and the addition of the forward index data are successful (S150).

FIG. 7 shows another example of the data input process according to the embodiment of the present invention.

When inverse index data having a combination of a field name and a field value of input data as a key is added to a database and there is inverse index data having a combination of a field name and a field value of the input data as a key, The department adds the identification value and the timestamp of the input data as the value items of the existing inverse index data. For example, in FIG. 7, input data having an employee number of '103', a name of 'donor', and an age of '26' is input. At this time, since the inverse index data having the key of "age = 26" is present in the database, the data input unit sets the combination of the employee number '103' and the time stamp '<ts4> We add it as a value item of data.

If the addition of the inverse index data is successful, the data input unit adds forward index data having the identification value of the input data as a key and having a combination of the field name, the time stamp, and the field value of the input data as the value items. This is the same as described in the example of Fig.

FIG. 8 shows another example of the data input process according to the embodiment of the present invention. 8, input data in which the employee number is "102" and the sex (field name) is "female" (field value) is input.

6 and 7, the data input unit stores the inverse index data having the combination of the field name and the field value of the input data as keys and the combination of the identification value and the time stamp of the input data as the value items, Add. That is, in the example of Fig. 8, the inverse index data having the combination of the employee number '102' and the time stamp '<ts4>' as a value with "gender = female" as the key is added to the database.

Then, the data input unit adds forward index data having the identification value of the input data as a key to the database. At this time, when forward index data having an identification value of input data as a key exists in the database, a combination of a field name, a time stamp, and a field value of the input data is added as a value item of the existing forward index data. That is, in the example of FIG. 8, since the forward index data having the employee number '102' as a key exists in the database, the combination of 'sex', 'ts 4' Add it as a value item.

9 shows an example in which two or more pieces of input data having the same contents are input. In the example of FIG. 9, the first input data having the employee number '103', the name 'doner', and the age '28' is inputted with the time stamp of '<ts 4>', The input data is entered with '<ts 5>' which is the latest time stamp than '<ts 4>'. That is, the first input data and the second input data having the same identification value, field name, and field value can be input with different time stamps. Such overlapping input of the same data may occur when a plurality of crawlers crawl and input the same data.

At this time, the reverse index data for the second input data may be input before the reverse index data for the first input data due to a network problem between the crawler and the data input unit or between the data input unit and the database. That is, when the inverse index data for the first input data is input, the database stores the inverse index data having a value item corresponding to the identification value of the first input data, May already exist in the database.

In this case, only when the time stamp of the first input data is newer than the time stamp of the value item of the existing inverse index data, the data input section sets the combination of the identification value of the first input data and the time stamp It is added as a value item of index data.

If the time stamp of the first input data is not more recent than the time stamp of the value item of the existing inverse index data, the data input unit stores the combination of the identification value of the first input data and the time stamp as the It returns the inverse index data addition failure information for the first input data without adding it as a value item. That is, in the example of FIG. 6D, the process of (3) is not performed. Also, in this case, since addition of the inverse index data to the first input data has failed, addition of the forward index data to the first input data does not occur.

FIG. 10 shows an example in which two or more pieces of input data having the same identification value are input. In the example of Fig. 10, the first input data having the employee number of '103', the name of 'doner' and the age of '28' is input with the time stamp of '<ts4>', , The second input data having the name 'Gogil-dong' and the age '30' is input with the latest time stamp '<ts 5>' than '<ts 4>'.

At this time, the forward index data for the second input data may be input before the forward index data for the first input data due to a network problem between the data input unit and the database. That is, when the forward index data for the first input data having the previous time stamp is inputted to the database with the same identification value as the identification value of the first input data, the field name of the first input data And forward index data having a value item corresponding to a combination of field values may exist in the database.

In this case, only when the time stamp of the first input data is newer than the time stamp of the corresponding value item of the existing forward index data, the combination of the field name, the time stamp and the field value of the first input data As a value item of the existing forward index data.

If the time stamp of the first input data is not more recent than the timestamp of the corresponding value item of the existing forward index data, the data input unit stores the combination of the field name, time stamp and field value of the first input data The forward index data addition failure information for the first input data is not added as the value item of the existing forward index data. That is, in the example of FIG. 10, the process of (4) is not performed.

<Update of data>

11 shows an example of the data update process according to the embodiment of the present invention. In the example of Fig. 11, the names and the ages of the employees whose employee number is '102' are updated to 'Gimchi-dong' and '30' respectively from 'Shim-chung' and '26'.

In the present invention, updating of data proceeds in the same process as addition of data. That is, the data input unit adds, to the database, the inverse index data having the combination of the field name and the field value of the input data as a key and the combination of the identification value and the time stamp of the input data as the value item (S120) If the addition is successful, the data input unit adds the forward index data having the identification value of the input data as a key and having a combination of the field name, the time stamp, and the field value of the input data as the value item (S130).

In the example of Fig. 11, with the key having the name = Gogil-dong as the key, the inverse index data having the combination of the employee number '102' and the time stamp '<ts4> Inverse index data having a value of '102' and a combination of a time stamp '<ts4>' is added to the database (S120). Then, having the employee number '102' as the key, and having the combination of 'name', 'ts 4>' and 'goddong' and the combination of 'age', 'ts 4' The forward index data is added to the database (S130).

That is, the data update can be completed only by adding the value item having the latest time stamp to both the inverse index data and the forward index data. At this time, since the value item having the previous time stamp is excluded from the search result by the later-described search method, the data is seen to be updated to the user even though the data is actually added. Also, a value item having a previous timestamp can be removed by garbage collection, which will be described later.

<Data deletion method>

FIG. 12 is a flowchart illustrating a method of deleting data in a database having a key-value structure according to an embodiment of the present invention, and FIG. 13 illustrates an example of a process of deleting data according to an embodiment of the present invention. In the example of FIG. 13, the age field of the employee whose employee number is '102' is deleted.

In the present invention, deletion of data is performed using addition of data. The data input unit receives the identification value and the field name of the data to be deleted (S210). Then, the data input unit adds a combination of the field name of the deletion target data, the time stamp, and the field value indicating the deletion target data to the forward index data having the identification value of the deletion target data as a key (S220).

At this time, the time stamp of the deletion target data is a value indicating the identification value of the deletion target data and the date and time when the field name is input to the data input unit. In addition, the field value indicating the data to be deleted may be a predetermined value, and may be null, for example.

If the step S220 is successful, the data input unit returns the data deletion success information (S230). If the step S220 fails, the data input unit returns the data deletion failure information (S240).

In this manner, the deletion process can be completed by merely adding a value item having a field name, a time stamp, and a preset value of deletion target data to the forward index data. At this time, the value item having the previous time stamp is excluded from the search result by the later-described search method. Therefore, even though the data is actually added, the data is shown to the user as deleted. In addition, the data input unit can input the identification value of the data to be deleted into the garbage queue, and the value item having the previous time stamp can be removed by the garbage collection described later.

In the above example, the identification value and the field name of the deletion target data are inputted. However, in the data deletion method according to another embodiment of the present invention, only the identification value of the deletion target data may be inputted . Thereafter, a value item having a field value (for example, null) indicating deletion target data may be added to all the field names included in the value items of the forward index data having the identification value of the deletion target data as a key.

<Data Search Method>

FIG. 14 is a flowchart illustrating a method of retrieving data in a database having a key-value structure according to an embodiment of the present invention. FIGS. 15A to 15B illustrate examples of retrieving data according to an embodiment of the present invention.

First, the data search unit receives a field name and a field value of search target data (S310). In the example of FIG. 15A, 'Age' is entered as the field name of the search object data and '30' is entered as the field value.

The data searching unit searches the database for the inverse index data having the combination of the field name and the field value of the search target data as a key (S320), and extracts the value item of the inverse index data (S330). In the example of FIG. 15A, the value item having the identification value '101' and the time stamp '<ts 1>' and the value item having the identification value '103' and the time stamp '<ts 5>' are extracted.

Then, the data searching unit searches the database for the forward index data having the identification value included in the extracted value item as the key (S340), and returns the forward index data as the search result (S350). In the example of FIG. 15B, the forward index data having the key '101' and the forward index data having the key '103' are the search results.

Here, in the example of FIG. 15B, there are two value items having 'age' as a field name of data to be searched in forward index data having '101' as a key. "Age <ts 1>: 30" and "Age <ts 2>: 26" are those. Here, "Age <ts 1>: 30" has a time stamp earlier than "Age <ts 2>: 26", so that it is known that it is updated old data. In the present invention, since the updated old data is deleted immediately after deletion using the garbage collection instead of immediately deleting the old data in the database, old data may exist in the database. In the example of FIGS. 15A to 15B, the forward index data having "101" as the key is searched based on "age = 30", but it should be excluded from the search result because it is searched by old data.

Therefore, only when the value item having the latest timestamp among the value items of the retrieved forward index data having the same field name as the field name of the retrieval object data has the same field value as the field value of the retrieval object data, The index data is returned as a search result (S350).

That is, the forward index data having '101' as the key has the field value of the value item "age <ts 2>: 26" having the latest time stamp among the value items having the & Is not the same as the value '26' and is excluded from the search result (S360). Therefore, only the forward index data having '103' as a key is included in the search result (S350).

If the value item having the latest time stamp among the value items having the same field name as the field name of the retrieval object data among the value items of the retrieved forward index data does not have the same field value as the field value of the retrieval object data, The data retrieval unit deletes the value item having the key of the forward data that is excluded from the search result among the value items of the extracted inverse index data as the identification value (S370). That is, in the example of FIG. 9C, since the value item having the identification value '101' and the time stamp '<ts 1>' among the value items of the extracted inverse index data is the old data to be deleted, 101 'and the time stamp' <ts 1> 'are deleted.

16A to 16B show another example of retrieving data. In the example of Fig. 16A, 'Age' is entered as the field name of the search object data and '26' is entered as the field value.

The data search unit searches the database for the inverse index data having the key "Age = 26", which is a combination of the field name and the field value, and extracts the value items "101 <ts2>" and "102 <ts3>" of the retrieved inverse index data do.

Then, the data search unit searches for forward index data having '101' as a key and forward index data having '102' as a key. In the example of FIG. 10B, the forward index data having '101' as a key includes two value items having a name as a field name and a value item having an age as a field name. Thus, a value item of a value item having the same field name that is not latest in time stamp is a value item to be deleted by garbage collection. Therefore, the data search unit returns only the value item having the latest time stamp among the value items having the same field name as the search result. Namely, only the value items of "Name <ts 2>: Hong Kil-dong" and "Age <ts 2>: 26" are included in the search result and "Name <ts 1>: Doodle", "Age <ts 1>: 30 "Is not included in the search result.

Figure 17 shows another example of retrieving data. The data search unit may additionally receive a range of identification values of the search target data. In the example of Fig. 17, '102 to 105' is entered as the range of the identification value of the search object data, 'age' is entered as the field name, and '26' is entered as the field value.

The data search unit searches the database for the inverse index data having the key "Age = 26 ", which is a combination of the field name and the field value. The retrieved inverse index data has "101 <ts2>" and "102 <ts3>" as the value items. At this time, the identification value included in the value item of "101 < ts2 >" is not included in the range of the identification value of the search object data.

Therefore, the data retrieval unit does not retrieve forward index data having '101' as a key, but retrieves only forward index data having '102' as a key in the database. The subsequent process is the same as described above.

< Garbage  How to Save>

FIG. 18 is a flowchart illustrating a method of garbage collection of data in a database having a key-value structure according to an embodiment of the present invention, and FIG. 19 illustrates an example of a garbage collection process according to an embodiment of the present invention.

The garbage collection unit checks whether there are two or more value items having the same field name for arbitrary forward index data (S410). A value item having the latest time stamp among the value items having the same field name is extracted (S420). The extracted value item becomes a garbage value item that is the object of garbage collection. In the example of FIG. 19, the value item of "Age <ts 3>: 26" among the forward index data having "102" as a key is a garbage value item.

The garbage collection unit searches the database for the inverse index data having the combination of the field value and the field name of the garbage value item as a key (S430), and stores the key of the forward index data for the garbage value item in the value items of the retrieved inverse index data as an identification value The branch value item is deleted (S440). In the example of Fig. 19, a value item having "102" as a key among the value items of the inverse index data having "age = 26" as a key is deleted.

Thereafter, the garbage collection unit deletes the garbage value item from the forward index data (S450). That is, in the example of FIG. 19, the value item of "Age <ts 3>: 26" is deleted from the forward index data having "102" as the key.

As described above, value items having duplicate field names can be deleted one by one in the above-described manner, and only the value items having the latest time stamp can be held.

On the other hand, in the example of FIG. 19, the field value of the value item having the latest time stamp among the value items having the 'age' as the field name is 'null'. That is, since the 'age' field is a field deleted by the above-described deletion method, a value item having 'null' as a field value is also subjected to garbage collection. Accordingly, the garbage collection unit deletes the value item having 'null' as the field value in the forward index data (S460).

In the above example, it is described that the garbage collection unit checks whether there are two or more value items having the same field name for arbitrary forward index data. However, garbage collection may be performed on the forward index data having the identification value stored in the garbage queue as a key.

The method of adding, updating, and deleting data in the database having the above-described key-value structure of the present invention is performed only by the additional operation of data internally and does not perform a process of internally modifying data, Update, and deletion of data without error such as collision with each other even if the addition, update, or deletion is requested.

In addition, when data is updated or deleted, substantial deletion of data is performed later using garbage collection or the like, and since the task is seen to have been completed when data is added to the user, There is a quick effect.

It will be understood by those skilled in the art that the present invention may be embodied in other specific forms without departing from the spirit or essential characteristics thereof. It is therefore to be understood that the above-described embodiments are illustrative in all aspects and not restrictive. The scope of the present invention is defined by the appended claims rather than the foregoing detailed description, and all changes or modifications derived from the meaning and scope of the claims and the equivalents thereof are included in the scope of the present invention Should be interpreted.

Claims (14)

A time stamp, and a time stamp of the data with a key as a key, a combination of a field name and a field value of data as a key and a combination of an identification value and a time stamp of the data as a value item, Value structure storing a forward index data having a combination of values as a value item, the method comprising the steps of:
(a) receiving a field name and a field value of search target data,
(b) searching in the database for inverse index data having a combination of a field name and a field value of the search target data as a key,
(c) searching the database for forward index data having an identification value included in a value item of the searched inverse index data as a key, and
(d) returning the searched forward index data as a search result, wherein a value item having a latest time stamp among the value items of the searched forward index data having the same field name as the field name of the searched data, And returning the retrieved forward index data as a retrieval result only when the retrieved forward index data has the same field value as the field value. The method according to claim 1,
The step (d)
And when only two value items having the same field name among the value items of the forward index data as the search result exist, including only the value item having the latest time stamp in the search result A method for retrieving data from a database having a value structure. The method according to claim 1,
The step (d)
If the value item having the latest time stamp among the value items of the retrieved forward index data having the same field name as the field name of the retrieval object data does not have the same field value as the field value of the retrieval object data, And deleting a value item having a key of the searched forward data as an identification value among the value items of the index data. The method according to claim 1,
The step (a)
Inputting a range of identification values of data to be searched,
The step (c)
Value structure having an identification value corresponding to a range of an identification value of the data to be searched as a key among identification values included in a value item of the searched inverse index data in the database. How to retrieve data from a database. A time stamp, and a time stamp of the data with a key as a key, a combination of a field name and a field value of data as a key and a combination of an identification value and a time stamp of the data as a value item, 1. A method of inputting data in a database having a key-value structure storing forward index data having a combination of values as a value item,
(a) receiving an identification value, a field name, and a field value of input data;
(b) inputting, into the database, inverse index data having a combination of a field name and a field value of the input data as a key and a combination of an identification value and a time stamp of the input data as a value item; and
(c) inputting the forward index data having the identification value of the input data as a key and having a combination of a field name, a time stamp, and a field value of the input data as a value item, when the step (b) Valued structure, wherein the key-value structure has a key-value structure. 6. The method of claim 5,
The step (b)
When a reverse index data having a combination of a field name and a field value of the input data is present in the database, a combination of an identification value and a time stamp of the input data is added as a value item of the existing reverse index data Wherein the key-value structure is a key-value structure. The method according to claim 6,
The step (b)
When there is inverse index data having a value item corresponding to the identification value of the input data in the database with a combination of a field name and a field value of the input data as a key, Adding a combination of an identification value and a time stamp of the input data as a value item of the existing inverse index data only when the time stamp of the value item is newer than the time stamp of the value item. How to enter data into a database with a value structure. The method according to claim 6,
The step (c)
Adding a combination of a field name, a time stamp, and a field value of the input data as a value item of the existing forward index data when forward index data having an identification value of the input data exists as a key in the database, Wherein the key-value structure is a key-value structure. 9. The method of claim 8,
The step (c)
When the forward index data having a value item corresponding to the combination of the field name and the field value of the input data is present in the database with the identification value of the input data as a key, Adding a combination of a field name, a time stamp, and a field value of the input data as a value item of the existing forward index data only when the latest entry is newer than the timestamp of the entry. A method for entering data into a database having a key-value structure. 6. The method of claim 5,
Returning data input success information if the step (c) is successful, and returning data input failure information if either step (b) or step (c) fails. A method for inputting data into a database having a key-value structure. A time stamp, and a time stamp of the data with a key as a key, a combination of a field name and a field value of data as a key and a combination of an identification value and a time stamp of the data as a value item, 1. A method for deleting data in a database having a key-value structure storing forward index data having a combination of values as a value item,
(a) receiving an identification value and a field name of data to be deleted, and
(b) adding, as a value item, a combination of a field name, a time stamp, and a field value indicating deletion target data of the deletion target data to the forward index data having the identification value of the deletion target data as a key A method for deleting data from a database having a value structure. 12. The method of claim 11,
And wherein the field value indicating that the data to be deleted is null is a method for deleting data in a database having a key-value structure. A time stamp, and a time stamp of the data with a key as a key, a combination of a field name and a field value of data as a key and a combination of an identification value and a time stamp of the data as a value item, 1. A method for garbage collection of data in a database having a key-value structure storing forward index data having a combination of values as a value item,
(a) when the forward index data in the database has two or more value items having the same field name, the key value of the field value and the combination of field names of the garbage value item having the latest timestamp among the two or more value items Retrieving the inverse index data from the database,
(b) deleting the value item having the key of the forward index data as the identification value from the searched inverse index data, and
(c) deleting the garbage value entry from the forward index data. &lt; Desc / Clms Page number 19 &gt; 14. The method of claim 13,
(d) if the field value of the value item of the forward index data is a field value indicating that the data item is to be deleted, deleting the value item having the latest time stamp Of garbage collection method.

Images