KR101705461B1 - Method and apparatus for encoding and decoding strings - Google Patents

Abstract

The present invention provides a dictionary based compressing method capable of being implemented on hardware. The present invention comprises: a step of searching for a partial string corresponding to a window of the entire strings in each of a plurality of dictionaries; a step of compressing the partial string when the partial string is searched in each of the plurality of dictionaries; and a step of registering the partial strings in any one of the plurality of dictionaries when the partial string is not searched in each of the plurality of dictionaries.

Description

[0001] METHOD AND APPARATUS FOR ENCODING AND DECODING STRINGS [0002]

The following embodiments relate to data compression and decompression methods, and more particularly, to a method of compressing and releasing data using a dictionary-based data management method.

Data compression techniques are processes that encode data such that the resulting representation has fewer bits than the original representation uses, i.e., the process of storing the data to occupy less space than usual. Compression techniques enable communication devices to transmit or store the same amount of data in fewer bits. The compression operation includes a predetermined encoding algorithm for receiving the original data and generating compressed data.

Data compression is widely used in backup utilities, spreadsheet applications and database management systems. Some types of data, such as bit-mapped graphics, can be compressed to a fraction of their original size through data compression.

Data compression is mainly divided into two types of compression: lossless compression and lossy compression. The lossless compression is reversible so that the original data can be reconstructed. On the other hand, a worst-case data compression scheme can achieve a higher compression rate, although some data loss may occur.

The lossless data compression scheme can be applied to data such as substrings, executable programs, and the like. A large amount of storage space can be saved through data compression. Data compression is achieved using a data compression algorithm. Several individual algorithms may be used depending on the type of data compression to be performed. Dictionary-based compression (or preshaped compression) corresponds to no-loss compression.

Data compression is possible by using various algorithms such as Huffman coding, arithmetic coding, dictionary based / substitution algorithm, and dynamically generated dictionary. The dictionaries can improve the data compression ratio with complex data types, frequent data changes, and / or data values with no obvious boundaries.

Software-implemented dictionary compression algorithms usually use a hash function to obtain the dictionary memory address. Common hash functions return a 32-bit or 16-bit hash value and use it as a memory address. However, in the case of a general PC, a memory corresponding to the size of an address converted through an MMU or a cache may not be required. However, when a software-implemented dictionary compression algorithm is fabricated with hardware, the system needs a memory corresponding to the size of the converted address in principle. A dictionary of software-based compression may require a single 4 GB memory space.

Even if you need 4GB of memory space, you can run the system with less than 4GB of physical memory using virtual memory, such as MMU. However, additional processing may be required in this case.

Therefore, when a compression system is implemented in hardware, if a pre-compression algorithm implemented by software is directly applied to the memory space without processing, very large memory is required. In addition, complicated and separate processing is required to reduce the required memory size.

As one embodiment of the present invention, there is provided a method of searching a plurality of dictionaries in a plurality of dictionaries, a method of compressing a string when being searched, and a method of registering dictionaries in advance when not searched, .

As an embodiment of the present invention, by providing a configuration for searching for a compressed portion, a configuration for extracting data corresponding to a compressed portion from a dictionary, and a configuration for replacing a compressed portion with extracted data, And provides a method for decompressing compressed data.

According to an embodiment of the present invention, there is provided a character string compression method comprising the steps of: searching a plurality of dictionaries for a partial character string corresponding to a window in an entire character string; when the partial character string is searched in each of a plurality of dictionaries, And registering the partial string in any one of the plurality of dictionaries when the partial string is not retrieved in each of the plurality of dictionaries.

In the character string compression method, the searching step may be a character string compression method of searching a plurality of partial strings extracted from the entire character string while moving the window, in each of the plurality of dictionaries.

In the string compression method, the searching step may be a string compression method for searching a plurality of dictionaries for a partial string starting from each character in the window.

In the string compression method, the searching step may be a string compression method of searching for the partial string in a plurality of dictionaries corresponding to the number of characters included in the window.

In the string compression method, the searching step may be a string compression method for searching the plurality of dictionaries for each of the plurality of dictionaries based on an address value obtained by converting a starting character in the partial string.

The character string compression method according to claim 1, wherein the registering step comprises: a step of registering, in a dictionary corresponding to a position of each character in the window among the plurality of dictionaries, And a string compression method for registering the position of the partial string with respect to the character string.

The character string compression method according to claim 1, wherein the step of registering further comprises the step of registering, when the position of the other partial string and the position of the other partial string are registered in the address value obtained by converting the starting character of the partial string in the dictionary corresponding to the position of each character in the window And determining whether to register the partial string or the other partial string by comparing the frequencies appearing in the entire string.

The method of claim 1, wherein, in the case of searching for a matching character string having a minimum length among the plurality of dictionaries among a plurality of dictionaries, And may be a string compression method that outputs the location of the substring.

A method for decompressing a character string according to an embodiment of the present invention includes the steps of searching for a compressed portion in a compressed character string with a matching length of a partial character string and a position of the partial character string with respect to the entire character string, Recovering a string of the compressed length from the corresponding string, and replacing the compressed portion with the recovered string.

As an embodiment of the present invention, it is possible to provide a configuration in which a character string is simultaneously searched in a plurality of dictionaries by a character string held by each dictionary, a configuration in which a character string is searched when it is searched, The compression throughput can be improved without processing.

As an embodiment of the present invention, by providing a configuration for searching for a compressed portion, a configuration for extracting data corresponding to a compressed portion from a dictionary, and a configuration for replacing a compressed portion with extracted data, The compressed data can be decompressed.

FIG. 1 shows an apparatus that implements a dictionary-based compression algorithm according to one embodiment in hardware.
2 illustrates a plurality of dictionary implemented forms on memory according to one embodiment.
FIG. 3 illustrates a window and a look-ahead window moving with respect to data to be compressed according to an embodiment.
Fig. 4 shows the relationship between a window and a plurality of dictionaries according to an embodiment, a structure of a character string memory and an offset memory constituting each dictionary, and a relationship between an ASCII code table and an address.
5 shows an algorithm for performing compression according to one embodiment.
6 shows a search step according to one embodiment.
7 illustrates a registration step according to one embodiment.
8 illustrates a compression step according to one embodiment.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

FIG. 1 shows an apparatus that implements a dictionary-based compression algorithm according to one embodiment in hardware.

The memory 130 may provide storage for a plurality of dictionaries. Each dictionary can consist of a pair of string memory and offset memory. The string memory stores the substring of the entire string to be compressed, and the offset memory stores the position of the substring in the entire string.

The processor 110 may perform a search step, a registration step, and a compression step on a substring whose starting point is the character string included in the window while moving the window from the entire character string input from the input / output unit 120. [ Here, the window may mean a range where the current search, registration, and compression steps are to be performed. Processor 110 may retrieve the substring from a dictionary stored in memory 130. [ If the substring is retrieved from the dictionary, the processor 110 may perform compression. The compressed result may be output through the input / output unit 120. If the substring is not retrieved from the dictionary, the processor 110 may register the substring in advance without performing compression.

The input / output unit 120 may receive the entire string to be compressed and transmit it to the processor 110. The input / output unit 120 may output an uncompressed portion and a compressed portion of the entire character string.

FIG. 2 illustrates a plurality of pre-implemented forms on memory 130 according to one embodiment.

The memory 130 may provide space for storing a plurality of dictionaries. Each dictionary may consist of a string memory 210 and an offset memory 220. The string memory 210 may be a storage space necessary for registering a partial string when a partial string starting from each character included in the window in the entire string to be compressed is not searched in a plurality of dictionaries. The string memory 210 may have a depth and a width. Where depth can refer to the address of the string memory.

The offset memory may provide the space required to store the location of the substring of the entire string. That is, if the partial string is not retrieved from the dictionary and the partial string is registered in the string memory 210 of the specific dictionary, the position of the partial string can be stored in the offset memory of the same dictionary. Offset memory can have depth and width. Where depth can refer to the address of the string memory.

FIG. 3 illustrates a window moving with respect to data to be compressed according to an embodiment.

The first window 301 may include a character string to be currently compressed among the entire character strings. The second window 302 may include a string to be compressed in the next section and may be directly attached to the first window 301. The second window 302 may be the same size as the first window 301. A simple "window" expression throughout the specification may refer to a first window.

Unlike the case of the compression algorithm implemented by software, in the case of compression implemented by hardware, since a partial string having the same length as the first window 301 must be stored in advance, it is necessary to know what partial string should appear next. Accordingly, the second window 302 may be required.

Processor 110 may perform the searching, registering, and compressing steps concurrently on the entire substring starting from each character in the first window 301. [ In this regard, the processor 110 may perform parallel processing. When the processor 110 completes the process for all substrings starting from each character in the first window 301, the first window 301 can move to the position of the second window 302, Window 302 may move to the next section. The first window 301 can move until the first window 301 does not contain a character. If the first window 301 does not contain a character, the compression process is terminated.

4 shows a correspondence relationship between a first window 301 and a plurality of dictionaries according to an embodiment, a structure of a character string memory 401 and an offset memory 402 constituting each dictionary, an ASCII code table 403, .

The position of each character in the first window 301 can correspond to each of a plurality of dictionaries. The number of characters included in the first window 301 may be the same as the number of dictionaries. For example, if the number of characters included in the window is N, the number of dictionaries may be N. [ The correspondence between the first window 301 and the dictionary is meaningful in the registration step. For example, in the registration step, the substring starting from the fifth character included in the window 301 can be registered in the fifth dictionary. In another embodiment, the number of dictionaries may be greater or less than the number of characters the window contains. In this case, the processor 110 may associate a plurality of dictionaries and positions of characters included in the window according to a preset reference.

The string memory 401 may have a depth and a width. Depth can refer to the address of the string memory. In an embodiment, when a value obtained by converting the first character of a substring into an address is used as an address, one alphabet character is one byte, and thus the depth may have an address of 0 to 255, that is, 28 addresses. A function that converts a character to an address may be by ASCII code conversion. If the maximum compression length of the character string is set to the number of characters that can be included in the first window 301, the width may be equal to the number of characters that the first window 301 can include. For example, if the number of characters that can be included in the first window 301 is N, the width of the character string memory 401 may also be N. [ Therefore, when the processor 110 compresses the partial string to the maximum, the processor 110 stores the partial string (length N) corresponding to the length of the number of characters included in the first window 301 in the string memory 401 It can be compressed to a string (length N).

The string memory 401 may store a string having a length equal to the number of characters included in the first window 301 in which compression is currently performed. Since the substring starting from the first non-first character in the first window 301 is shorter than the length of the first window 301, the processor 110 can retrieve and store the characters of the second window 302 have. That is, the processor 110 may store a substring of a length equal to the width of the string memory 401. [

A software - based dictionary - based compression algorithm uses a hash function for address translation. At this time, if the bucket of the hash table is 32 bits, 232 memory space is required. Since a very large amount of memory is required when implementing the same in hardware, an embodiment of the present invention implemented by hardware can use a simple hash function rather than a compression algorithm implemented by software.

The processor 110 may implement a simple hash function using the ASCII code table 403. [ The processor 110 may use the number derived from the ASCII code table 403 of the first character of each substring to be compressed as an address value. Since the ASCII code has a value of 0 to 255, the number derived according to the ASCII code table 403 may correspond to the depth of the character string memory 401 or the offset memory 402 of FIG.

The offset memory 402 may have a depth and a width. The depth of the offset memory 402 may refer to the address in the same manner as the depth of the string memory 401. The width stores the location of the substring, so it may vary depending on the length of the entire string. If the length of the entire string is 256, the offset memory 402 must be able to store 256 positions, so the width may be 1 byte. That is, K may be 8.

5 shows an algorithm for performing compression according to one embodiment.

The processor 110 may move the window 301 from the entire character string to the entire process for the partial character string starting from the character string in the window 301. [ The window 301 may refer to the first window 301. At step 510, the processor 110 may search for each of the substrings starting from each character in the current window 301 in a plurality of dictionaries. Searching in a plurality of dictionaries can be performed simultaneously for all of the substrings. Accordingly, the processor 110 may perform the parallel processing in units of the windows 301. At this time, the processor 110 may compare the string stored in the string memory of the plurality of dictionaries with the partial string based on the converted address value of each character.

The dictionary-based compression algorithm implemented by the software reads the values in the input buffer one byte at a time, and directly scans the window 301 one byte at a time, thereby processing it in a serial manner. However, according to the embodiment of the present invention, the dictionary-based compression algorithm implemented in hardware can perform the parallel processing because the searching process is performed in a plurality of dictionaries simultaneously for the whole substrings starting from each character included in the window 301 Can be performed. For example, if the window 301 is 8 bytes, the processor 110 may perform parallel processing because it compares 8 subsequences starting with 8 characters to a plurality of dictionaries at the same time.

The length of the substring may be the same as the length of the window 301 (or the first window). At this time, characters of a short length can be taken from the second window 302 and supplemented.

At step 520, the processor 110 may determine the matching length of the substring to compress and the string searched in the dictionary (string memory 210). Processor 110 may perform compression if the match length is greater than or equal to a predetermined minimum length. In one embodiment, since the offset and the matching length usually occupy 3 bytes, it is necessary to match 4 bytes or more so that the gain can be 1 byte. Therefore, it is necessary to match 4 bytes or more to perform compression. Where the minimum length may be 4 bytes.

At step 530, the processor 110 may perform compression of the substring. When the substream to be currently compressed is registered in the dictionary (string memory 210), the processor 110 determines that the current compression should be performed with a match length that matches the offset of the string You can substitute a string.

The offset of the partial string to be compressed can be obtained by subtracting the value read from the offset memory 220 at the position of the window 301. In other words, the processor 110 obtains the relative position of the partial string and the same string registered in advance (the string matching the minimum length or more) instead of the position (absolute position) where the same string . Since the relative position is generally smaller than the absolute position, the compression effect can be increased when the relative position is output as the compression result.

If the match length is less than the minimum length, the processor 110 may determine whether the substring has not been retrieved from the dictionary. In other words, the processor 110 may determine whether the matching length of the partial string and the string registered in the dictionary (string memory 210) is zero. If the matching length is 0, the partial string is not registered in the dictionary. Therefore, the processor 110 may register a partial string having a matching length of 0 in advance to compress the partial string to be searched next. At this time, when the position of the dictionary and address to be registered overlap with the already registered string, a string having a higher frequency can be registered based on the appearance frequency.

In another embodiment, the processor 110 may determine whether the matching length of the substring and the string registered in the dictionary (string memory 210) is less than the minimum length. If the matching length is less than the minimum length, the substring can be registered in advance. At this time, if the dictionary and address to be registered overlap in step 540, the substream may be registered in advance by comparing the appearance frequency of the already registered string including the matching part with the frequency of occurrence of the substring .

In step 550, the processor 110 may look up a dictionary corresponding to the position of the character contained in the window 301 among the plurality of dictionaries. That is, if the processor 110 wishes to register a substring starting from a specific character contained in the window 301, the processor 110 determines whether the first character (a specific character) of the substring is the first character And then register the partial string in the dictionary of the same sequence number. In another embodiment, the processor 110 may determine in which dictionary of the plurality of dictionaries to register a substring according to a preset criteria. In another embodiment, the processor 110 may determine at which address in the dictionary string memory to register the substring according to a preset criteria if the dictionary is selected. In addition, the processor 110 can find an address to register the substring based on the address value that has converted the start character of the substring. The processor 110 may register the partial string in the string memory 210 based on the corresponding address of the found dictionary and register the position of the partial string in the offset memory 220 with respect to the entire string.

When the search, registration, and compression for the substring starting from each character in the window 301 are all completed, the window 301 can move to the next section in step 560. At step 570, the processor 110 may determine if there are no characters remaining in the window 301. [ If there are no remaining characters, the processor 110 may determine that compression has been completed for the entire string. If there are any remaining characters, the processor 110 may repeat the entire process from step 510.

6 shows a search step according to one embodiment.

The entire string 610 may be divided into three sections. The first window 301 may be located in the first section. At this time, partial strings 620 having the same length as the first window 301 starting from each character included in the first window 301 can be searched. The processor 110 searches for a character string having a length equal to the length of the number of characters included in the window so that the missing character can be supplemented through the characters included in the second window 302. [ For example, in the case of HABCDEYU, the processor 110 can supplement ABCDEYU from the second window 302 because it is only H in the first window 301. [

The character string memory 630 and the offset memory 640 may have the same width as the number of characters included in the first window 301. In the search step, each substring 620 can be compared with a string stored in the corresponding address of all the string memory 630 at the same time. In other words, because the search is made simultaneously for the entire substring 620, the processor 110 can perform parallel processing. Here, the address may refer to a value obtained by address-converting the first character of each substring. For example, since A is converted to 97 in the ASCII code table 403 of FIG. 4, the processor 110 can compare ABCDEFGH with the string stored at address 97 of the string memory 630.

The processor 110 may determine the matching length of the string stored in the substring 620 and the string memory 630. If the match length is greater than or equal to the minimum length, the compression step 530 of FIG. 5 may proceed. If the match length is less than the minimum length, then the registering step 550 of FIG. 5 may proceed if there is no matching length, i. E., A string not previously registered as a non-searched string.

If there is a match length but less than the minimum length, the processor 110 can output it without performing compression. This is because there is no gain in compression when compressing substrings less than the minimum length.

7 illustrates a registration step according to one embodiment.

In FIG. 7, the first window 301 is located in one section. The processor 110 may determine whether the substring has not been retrieved from the dictionary and may perform the registration step if it is not retrieved (or the match length is zero). The number of characters included in the first window 301 may be the same as the number of dictionaries and the position of each character included in the first window 301 may correspond to each of the plurality of dictionaries.

The processor 110 can find a dictionary corresponding to a position of a character included in the first window 301 among a plurality of dictionaries. The processor 110 may also find the address based on the address value that translated the start character of the substring. The processor 110 may register the partial string in the string memory 630 based on the corresponding address of the found dictionary and register the position of the partial string in the offset memory 710 with respect to the entire string.

For example, when the processor 110 registers BCDEFGHA in FIG. 7, among the characters included in the first window in the section 1, B is the second character, so it can correspond to the dictionary 1 which is the second dictionary. The processor 110 may register the substring BCDEFGHA at the address of the dictionary corresponding to 98, which is the result of the conversion by the ASCII code table of B, which is the first character of the substring.

In the dictionary corresponding to the position of each character included in the first window 301, the string memory 630 corresponding to the address value obtained by converting the start character of the partial string, and the other partial strings in the offset memory 710, If the position of the substring has already been registered, a problem of overwriting may occur. In this case, the processor 110 can determine which one to register by comparing the frequency of appearing in the entire string with the partial string already registered in the entire character string and the partial character string to be registered at present.

8 illustrates a compression step according to one embodiment.

The first window 301 may be located in the current section 2 after ABCDEFGH is registered at the address 97 of the dictionary 0 in the first section. The registered substring ABCDEFGH containing the ABCDE in which the actual compression of the substring ABCDEYUI is actually registered is registered at the address 97 of the dictionary 0. When the minimum length of compression is set to 4 bytes, the processor 110 compresses ABCDE since the length of the matching string (5 bytes) is equal to or more than the minimum length (4 bytes) of the partial string ABCDEYUI and the registered partial string ABCDEFGH can do.

The processor 110 may replace a character string (ABCDE) to be actually compressed with a relative position offset in which a character string (ABCDE) to be actually compressed is present. The position offset can be obtained by subtracting the value read from the offset memory 640 from the position of the first window 301. Here, the value read from the offset memory 640 may indicate the position where the partial string registered in the string memory of the same address appears in the entire string.

The position of the first window 301 may indicate the position of the first character included in the first widow 301. In FIG. 8, if the position of each character is ABCDEFGH and the position of A is 0, the position of the first window 301 may be 8. The value read from the offset memory 640 may be zero. Therefore, the relative position offset of the string (ABCDE) can be 8 - 0 = 8.

If the substrings starting from each character in the first section are not searched in the dictionary and if neither of them is compressed, the result of processing from the section A to the substrings starting from A may be the same as the result 820. In the result (820), ABCDEFGH is the output of the character string of the first section as it is not compressed, and 85 is the result of ABCDE of ABCDEYUI being compressed, where 8 is the position offset and 5 means the matching length (ABCDE) can do.

According to another embodiment of the present invention, it is possible to provide a method of releasing a compressed string by a dictionary-based compression algorithm implemented in hardware.

The processor can retrieve the compressed portion of the compressed string by matching the length of the substring and the position of the substring for the entire string. The compressed string can be recovered from the string corresponding to the position of the compressed portion. The processor can complete decompression by replacing the compressed portion with the recovered string.

The apparatus described above may be implemented as a hardware component, a software component, and / or a combination of hardware components and software components. For example, the apparatus and components described in the embodiments may be implemented within a computer system, such as, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate array (FPGA) , A programmable logic unit (PLU), a microprocessor, or any other device capable of executing and responding to instructions. The processing device may execute an operating system (OS) and one or more software applications running on the operating system. The processing device may also access, store, manipulate, process, and generate data in response to execution of the software. For ease of understanding, the processing apparatus may be described as being used singly, but those skilled in the art will recognize that the processing apparatus may have a plurality of processing elements and / As shown in FIG. For example, the processing unit may comprise a plurality of processors or one processor and one controller. Other processing configurations are also possible, such as a parallel processor.

The software may include a computer program, code, instructions, or a combination of one or more of the foregoing, and may be configured to configure the processing device to operate as desired or to process it collectively or collectively Device can be commanded. The software and / or data may be in the form of any type of machine, component, physical device, virtual equipment, computer storage media, or device , Or may be permanently or temporarily embodied in a transmitted signal wave. The software may be distributed over a networked computer system and stored or executed in a distributed manner. The software and data may be stored on one or more computer readable recording media.

The method according to an embodiment may be implemented in the form of a program command that can be executed through various computer means and recorded in a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, and the like, alone or in combination. The program instructions to be recorded on the medium may be those specially designed and configured for the embodiments or may be available to those skilled in the art of computer software. Examples of computer-readable media include magnetic media such as hard disks, floppy disks and magnetic tape; optical media such as CD-ROMs and DVDs; magnetic media such as floppy disks; Magneto-optical media, and hardware devices specifically configured to store and execute program instructions such as ROM, RAM, flash memory, and the like. Examples of program instructions include machine language code such as those produced by a compiler, as well as high-level language code that can be executed by a computer using an interpreter or the like. The hardware devices described above may be configured to operate as one or more software components to perform the operations of the embodiments, and vice versa.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed exemplary embodiments. For example, it is to be understood that the techniques described may be performed in a different order than the described methods, and / or that components of the described systems, structures, devices, circuits, Lt; / RTI > or equivalents, even if it is replaced or replaced. Therefore, other implementations, other embodiments, and equivalents to the claims are also within the scope of the following claims.

301: First window
302: second window
401: String memory
402: offset memory
403: ASCII code table

Claims (9)

Retrieving a substring corresponding to the window in the entire string in each of the plurality of dictionaries;
When a matching portion of at least one dictionary registered in the plurality of dictionaries and the partial string is searched for, the length of the matching portion and the position of the matching portion with respect to the entire character string are used, Compressing the matching portion within a string;
Registering the partial string in any one of the plurality of dictionaries when the partial string is not searched in each of the plurality of dictionaries as a match of a predetermined minimum length or more,
A string compression method that includes.
The method according to claim 1,
Wherein the searching comprises:
And searching the plurality of dictionaries for each of the plurality of partial strings extracted while moving the window in the entire string.
The method according to claim 1,
Wherein the searching comprises:
And a substring starting from a character corresponding to each position in the window is searched in each of the plurality of dictionaries.
The method according to claim 1,
Wherein the searching comprises:
And searching for the substring in the plurality of dictionaries corresponding to the number of characters included in the window.
The method according to claim 1,
Wherein the searching comprises:
And searching for each of the plurality of dictionaries in the plurality of dictionaries based on an address value obtained by converting a start character in the partial string.
6. The method of claim 5,
Wherein the searching comprises:
And searching said plurality of dictionaries for each of said plurality of dictionaries based on an address value obtained by converting a start character in said partial string by an ASCII code table.
The method according to claim 1,
Wherein the registering step comprises:
The position of the partial character string for the entire character string and the position of the partial character string for the entire character string based on the address value obtained by converting the starting character of the partial character string into a dictionary related to the position of the character corresponding to each position in the window among the plurality of dictionaries A string compression method that registers.
8. The method of claim 7,
Wherein the registering step comprises:
When a position of another partial string and a position of another partial string are registered in an address value obtained by converting a starting character of the partial string in a dictionary corresponding to a position of a character corresponding to each position in the window, And determining which of the partial string and the other partial string is to be registered.
Retrieving a compressed portion in a compressed substring using a length of a matching portion of a plurality of pre-registered strings and a pre-compressed substring and a position of the matching portion with respect to the entire string;
Restoring the compressed substring of the length of the matching portion using the plurality of pre-registered strings corresponding to the position; And
Replacing the compressed portion with the recovered string,
The compressed substring may be a sub-
The pre-compressed substring corresponding to the window in the entire string is searched in each of the plurality of dictionaries, and a matching portion of at least one dictionary registered in the plurality of dictionaries with the pre- Compresses the matching portion in the pre-compressed substring using the length of the matching portion and the position of the matching portion with respect to the entire character string, If the partial string before compression is not searched for a match of at least a predetermined minimum length, the partial string before compression is registered by being registered in any one of the plurality of dictionaries
How to uncompress a string.

Images