JPH03104421A - System and device for data compression and data decoder - Google Patents

Abstract

PURPOSE:To improve the data compression rate and to enhance the utilizing efficiency of a recording medium by compressing a data coincident with a data string constituting a prescribed data block into a data string replaced with information of the number of data coincident consecutively, and identification information representing the location of the information of a data number in the replaced data string. CONSTITUTION:When a data of a 1st block is inputted entirely to a character compression circuit 6 repetitively, a control circuit 10 switches a signal line 17 and an output of a selector 5 is used as a data from a selector 18. When a data string of a 2nd block is inputted, a processing 9 compares it with an input data string of the 1st block stored in a buffer memory 3. When they are coincident, a signal line 12 is valid and an input data string of the 1st block and the same data of the input data string of the 2nd block inputted to a same data counter 13 generate the information of consecutive length. The data count inputted to the compression circuit 6 is compressed, and written in a buffer memory 8 through a selector 7.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a data compression method, a data compression device, and a data decompression device.

[Prior Art] When recording a data group consisting of a plurality of divided data blocks on a magnetic tape or the like, the method shown in FIG. 4(a)
As shown in the figure, recording the data blocks A to C themselves would result in poor recording efficiency on tape, etc., so conventionally,
The amount of data is compressed by using the correlation of continuous data.

The technology described in Japanese Patent Application Laid-Open No. 60-35373 improves the usage efficiency of recording media by using a repeated character compression method that allows compressed data to be restored regardless of whether the data is read in the forward or reverse direction. It is something that makes you

In this technique, data compression is performed for each data block, and when data compression is not performed, data blocks A to D are recorded as shown in FIG. 4(a). As shown in b), a new data block with a = d is recorded on the magnetic tape. The IBG (inter-data block gap) in the figure is provided between data blocks so that the data blocks can be separated.

Note that the effect of such data compression is that the stronger the autocorrelation of data within a block, the higher the compression rate.

[Problems to be Solved by the Invention] However, the above-mentioned conventional technology is insufficient in terms of data compression efficiency.

In many cases, each data block within a file consisting of multiple data blocks has the same number of bytes.
This is because, although it is known from experience that each data block is similar, conventional techniques do not utilize the correlation that exists between data blocks for data compression.

The present invention provides a data compression method, a data compression device, and a data restoration device that can alleviate the drawbacks of such conventional techniques, further improve the data compression rate, and improve the usage efficiency of recording media. It's about doing.

[Means for Solving the Problem] In order to achieve the above object, a compression method for a plurality of data blocks is provided, in which a data string constituting a data block other than a predetermined data block is sequentially compressed into the predetermined data. A data string in which the data that matches the data string that determines the block is replaced with information about the number of consecutively matching data,
A data compression method is provided, characterized in that the information is represented by identification information indicating the position of the information of the number of data in the replaced data string.

Further, in order to achieve the above-mentioned object, the present invention individually separates a data string of a predetermined data block and a data string of data blocks other than the predetermined data block within a predetermined data group consisting of a plurality of data blocks. means for comparing the data, a counting means for counting the consecutive number of identical data when the comparison results are the same, and a new data string by replacing the consecutive identical data with the count counted by the counting means. means for creating and a predetermined control information area;
means for setting a control bit corresponding to a count value position in the new data string, and compression corresponding to a data block other than the predetermined data block using the new data string and control information that is information in a control information area. A data compression device is provided, comprising means for creating a data block and means for creating a compressed data group from a compressed data block and the predetermined data block.

In addition, the present invention continuously matches data blocks other than a predetermined data block in a data group consisting of a plurality of data blocks with data that matches a data string constituting the predetermined data block. a compressed data block consisting of a data string replaced with a count value that is the number of data that has been replaced, and control information in which a control bit corresponding to the position of the count value in the replaced data string is set; and the predetermined data block. and, by referring to the control bits of the control information of the compressed data block in the compressed data group containing A data restoration device is provided, characterized in that it has means for replacing the data extracted by the count value from a position corresponding to a data position of a data block to be restored in a column.

Note that the data compression device may further include means for compressing the compressed data group using a predetermined coding rule, such as repeated character compression or Huffman code compression.

[Function] According to the data compression method according to the present invention, a data string constituting a data block other than a predetermined data block is sequentially concatenated with data that matches a data string constituting the predetermined data block. The information is compressed into a data string replaced with information on the number of data that matches, and identification information indicating the position of the information on the number of data in the replaced data string.

Further, according to the data compression device according to the present invention, the data string of a predetermined data block and the data string of data blocks other than the predetermined data block in a predetermined data group consisting of a plurality of data blocks are individually separated. Compare the data, and if the comparison results are the same, count the number of consecutive identical data. In addition, when a new data string is created by replacing continuous identical data with count values, both
A control bit corresponding to the count value position in the new data string is set to 1-.

Then, a compressed data block corresponding to a data block other than the predetermined data block is created using the new data string and control information that is information in the control information area, and further, a compressed data block corresponding to a data block other than the predetermined data block is created. and create a compressed data group.

Further, according to the data restoration device according to the present invention, the count value in the replaced data string is determined by referring to the control bit of the control information of the compressed data block. of the data string of the block,
Replace with the data extracted by the count value from a position corresponding to the data position of the data block to be restored.

In the data compression device, the compressed data group may be further compressed using a predetermined code rule such as repeated character compression or Huffman code compression.

As described above, if the predetermined amount of one data group is limited, even if only the third block and subsequent blocks are to be restored, a large amount of time is not required to restore the first block. Restoration can also be done quickly.

In addition, when recording compressed data on magnetic tape, there are generally many sequential reads, and random reads like the one mentioned above are rare, so the case of restoring data only from the third block onward is unlikely to occur. .

Furthermore, when recording compressed data on a magnetic tape, data that does not exist as compressed data (identification marks) is used instead of dividing it by non-data areas such as gaps between blocks as in the conventional technology. The gap between blocks may be substituted to further improve the usage efficiency of the recording medium. In this case, it is desirable to provide a conventional non-data area such as a gap for each data group including a plurality of blocks.

(Margin below) [Example] An embodiment of the present invention will be described below.

FIG. 1 shows the configuration of a data compression device according to this embodiment.

In the figure, 3 is a buffer that holds the data block of the data group, 5 is a selector that selects and outputs the input data IN2, the data from the selector 18, and data "0", and 6
is a repeating character compression circuit that repeatedly compresses data from selector 5, 7 is a selector for inserting an inter-block identification mark M, 8 is a buffer for holding output data, 9
13 is a comparator that compares the data block extracted into the buffer 3 with the input data block; 13 is an identical data counter that counts the number of consecutive identical data from the comparison result of the comparator 9;
14 is a control bit generator that generates a control bit to be described later;
18 is a selector that selects and outputs either the count result of the same data counter, the control bit generation, or the generated control bit and the manual data IN2; 10 is a control circuit that controls the entire control bit generation, etc.

The operation of this data compression device will be explained below.

As an example, this data compression device uses an input data string INI
It receives input from the host device, etc., such as signal 2, byte-based data transfer instruction signal REQI signal 1, and block-based data transfer end signal ENDI signal, compresses the data, and outputs the output data string OUT, etc. .

First, when the input data string of the first block of a predetermined data group is input from the bus line 2 in synchronization with the REQI signal 1, the input data string is processed when the ENDI (data transfer end signal) 4 is valid. The data is sequentially stored in the buffer memory 3 in 1-byte units until the number of data is reached.

At the same time, the input data of the first block passes through the selector 5, is input to the repetitive character compression circuit 6, is compressed, passes through the selector 7, and is stored in the buffer memory 9.

Next, when all of the data of the first block is input to the repetitive character compression circuit 6, the control circuit 10 switches the signal line 17 and outputs the output of the selector 5 from the selector 18 in order to process the second and subsequent blocks. Data.

Furthermore, when all the input data strings of the first block have been processed by the compression circuit 6 and stored in the buffer memory 8, the control circuit 10 controls the input data string between the blocks in order to clarify the boundary with the input data string of the second block. Data M called identification mark is 2
The byte is sent from the selector 7 to the buffer memory 8. This inter-block identification data M is 2-byte continuous data that can never appear in a data string after compression by repeated character compression, thereby making it possible to clearly indicate the boundaries between blocks.

Next, when the data string of the second block is inputted, this input data is compared with the input data string of the first block stored in the buffer memory 3 in the comparator 9.

At this time, the buffer memory 3 is switched from the write state to the read state by the control circuit 10 through the signal line 11, and the memory address is initialized by the END1 signal 4 indicating the end of the block, and then the buffer memory 3 is switched from the write state to the read state by the control circuit 10 through the signal line 11. Sequential addition is performed by REQ signal 1. As a result, the comparator 9 compares the input data string of the first block on the bus m23 with the input data string of the second block of the bus line 2. If so, enable signal line 12. This signal line ■2 is the control circuit 1
0 and the same data are input to the counter 13.

Since the same data counter 13 generates length information in which the same data of the input data string of the first block and the input data string of the second block are continuous, the same data counter 13 generates length information in which the same data of the input data string of the first block and the input data string of the second block are continuous. The counter adds up in response to the REQ signal 1, and outputs the counting result to the bus 15 as a data count value.

On the other hand, in the control circuit 10, when the signal line 12 changes from valid to invalid, the selector 18 is switched by the signal line 16, and the data count value on the bus line 15 is guided to the compression circuit 68.

The data count value input to the compression circuit 6 is compressed and written to the buffer memory 8 through the selector 7.

Further, the control circuit 10 sends a pit set signal to the control bit generation circuit 4 through the signal line 19 in order to generate a control bit indicating that the same data count value has been output as data.

The control bit generation circuit 14 takes in the signal line 24 from the control circuit 10 at the timing of the signal linework 9.

The signal line 24 is the bus [+j if 27 is the same data count value
This is a signal set to IN.

By the way, the counter width of the same data counter 3 is capable of expressing numerical values up to 8 bit width (0 to 255) in this example, and in the event that the same data counter 13 overflows, the carry signal 20 By outputting , the same data count value is once sent to the compression circuit 6, and after the count value is cleared by the reset signal 2 from the control circuit 10, it is counted again. Of course, when outputting the same data count value by carry, due to control bit generation,
The control circuit 10 has a signal line 19 and a signal line 24 connected to each other.
·do.

On the other hand, if the comparison result between the bus line 2 and the bus line 23 by the comparator 9 is inconsistent and the signal line 12 is invalid, the control circuit IO clears the signal line 24 in response to the inconsistent data. , set the signal line l9. As a result, u O 7+ is taken into the control bit generation circuit 14.

At the same time, the control circuit 10 switches the selector 18 using the signal line 16 and guides the input data of the second block to the compression circuit 6 via the bus line 2.

Thereafter, while the signal fil2 is invalid, the control circuit 10 causes the control bit generation circuit 14 to take in the control bit II O u for each byte of mismatched data.

By the way, as the comparison of the data strings progresses in the comparator 9, the register in the control bit generation circuit 4 becomes full. This is communicated to the control circuit 10 by enabling the FULL signal 22, and the control circuit 10 switches the selector 18 through the signal line 16 and sends the control bit string (8 bits) to the compression circuit 6.

As a result, the control bit string is inserted once every 8 bytes in the data strings sent to the compression circuit 6 for the second and subsequent blocks. FIG. 3 illustrates this situation.

According to this, the control bits I~ corresponding to the same data count numbers 81, 81a and 8lb are set to "1" and added (82, 82a) every 9th byte.

By the way, “OO” data 2 added to area 83
A byte is added when the data ends with less than 8 bytes in order to ensure that the byte data including the control bit comes every 8 bytes, and in this case, the corresponding control bit is always II I IT. By doing so,
When restoring, this unnecessary data can be separated.

Note that when adding this 11 0 0 I+ data, the control circuit 10 sets the signal line 24, and the signal wire operator 9 sets the control bit, and at the same time switches the signal line 17 so that the selector 5 outputs "OO" data. The data is transferred to the compression circuit 6 at the timing of the signal line 28.

This processing of data less than 8 bytes is continued until the FULL signal 22 from the control bit i/generation circuit 14 becomes valid, and at the time the FULL signal 22 becomes valid, the control circuit 10 connects the signal lines 16 and 17. Data from the switching bus line 25 is guided to the compression circuit 6.

After each process of the input data blocks described above is completed, the control circuit 10 outputs the inter-block identification data.
selects the inter-block identification mark M with the signal line 26,
Two bytes are sent to the buffer memory 8 in synchronization with the signal line 29.

Further, the processing up to this point is performed for the input data strings of the third block, the fourth block, etc. until the buffer memory 8 reaches a predetermined amount of data.

When the buffer memory 8 reaches a predetermined amount of data, the buffer memory 8
The control circuit is notified of this via the No. 3 line 30. In response to this, the control circuit 1o outputs the data in the buffer memory 8 up to that point to the bus line 31 as a new data group of one block. After the END1a37 signal is enabled, the above process is repeated if there is next input data N2.

By the way, since a magnetic tape recording device or the like adds the IBG to recorded data using the END signal, by validating and outputting the ENDI a signal 37 instead of the ENDI signal 4 as described above, the IBG is It will be added to each data group, not to each data block. Further, the control bits may not be inserted every 8 bytes, but may be inserted all at the end of the data group.

Furthermore, the compression circuit 6 is not limited to one that specifically performs repeated character compression, but can be similarly implemented and effective with other types of compression. Note that signals 4132, 32, and 34, which did not appear in the above explanation, are data transfer refusal signals, and signal lines 35, 36 are transfer instruction signals.

Next, a data restoration device according to this embodiment will be explained.

FIG. 2 shows the configuration of the data restoration device according to this embodiment.

In the figure, 43 is a buffer that holds the compressed data group to be restored, 44 and 49 are repetitive character compressed data restorers, 45 is a buffer that holds the l-th block data restored by the restorer 44, and 46 is a buffer that holds the compressed data group to be restored. A control circuit 48 that controls the overall operation is a selector that switches between data from the buffer 45 and data from the control circuit.

The operation of this data restoration device will be explained below.

For example, this data restoration device receives an input data string IN2 signal 42 from a recording device or the like, a data transfer instruction number a REQ2 signal 41 . Receives input from the host device, etc. from the recording device, etc., such as the END2 signal, which is the data transfer end signal in block units, restores the data compression, and outputs the data string O.
It was assumed that UT, etc. would appear.

First, one data group including a plurality of compressed data blocks is transmitted to the bus line 42 in synchronization with the REQ2 signal 41.
Input from

The input data group is all stored in the buffer memory 43, and at the same time, the data portion corresponding to the first block (the data string until the inter-block identification mark is detected for the first time) is stored in the restoration circuit 44. After the data is restored, it is stored in the buffer memory 45.

Thereafter, when output of the data string corresponding to the first block is requested, the control circuit 46 controls the selector 48 via the signal line 47 to selectively output the data in the buffer memory 45.

Also, if the second or subsequent blocks are requested,
The data of the necessary block is sent from the buffer memory 43 to a restoration circuit 49 of the same type as the restoration circuit 44, restored, and input to the control circuit 46.

The control circuit 46 performs data restoration processing for the second and subsequent blocks using 9 bytes as one unit.

The data input to the control circuit 46 has a format in which a 1-byte control bit string is added next to the 8-byte data shown in FIG.

The control circuit 46 controls the control bit string of the 9th byte (see FIG.
Byte data (81, 81a, 8lb and 8
3) are considered to be the same data count value, and the control circuit 46
operates the buffer 45 through the signal line 5o so as to output from the buffer memory 45 a number of data corresponding to the count value from the data position, and the selector 4 through the signal gland 47.
8 is controlled to selectively output the bus line 51.

Further, the control circuit 46 outputs the data bytes (84.84a, 84b and 84c) corresponding to the part where the bits of the 9th byte of data are 11 0 17 as is from the data taken into the control circuit 46. Thus, the selector 48 is switched by the signal line 47 to selectively output the bus 852 as the OUT signal 65.

When the above processing is completed for 8 bytes, the control circuit 46 receives 9 new bytes from the restoration circuit 49, and repeats the above processing for the second block and subsequent blocks in the same manner for one block.

If the control bit is It I I+ at the end of the data block indicated by the inter-block identification mark, indicating that the corresponding data bytes have the same data count value, but the count value is zero, The data shown in FIG. 3 83 indicates data added as dummy data to the remaining data bytes when the data length was not divisible by 8 bytes during compression. Delete.

When the restoration of one data block is completed, the control circuit 46 outputs the END signal 53.

Furthermore, the processing up to this point is performed in the buffer memory 4.
3 until it becomes empty, all data blocks in one group of data that have been read are restored.

Note that since the necessary block position can be detected by counting the END signal 54 and the inter-block identification marks detected within the control circuit 46, it is possible to restore any data block. In this case, after reading one data group into the buffer memory 43, the restoration of the first block is completed and recorded into the buffer memory 45,
Thereafter, the specified data block is detected by the means described above, and the corresponding data block is restored using the same means as described above.

Furthermore, the same restoration as described above is also possible in the reverse direction based on the control bits.

In addition, during the explanation of the above restoration process, the signal line 5 which is not explained
5, 56, 57 and 58 are valid data transfer signals;
Signal lines 59, 60, 6l and 62 are data transfer rejection signals, and signals B63 and 64 are buffer memory REA signals.
This is a D/WRITE switching signal.

As explained above, according to the present embodiment, especially in a data storage device such as a magnetic tape that performs sequential writing and sequential reading of data, the second block is The subsequent data is based on the data of the 1st eblock, and includes information on the data byte length of the part where the data matches, information on the mismatched data, and a control byte that indicates whether it is the byte length information or the mismatched data. By compressing information into data, the compression rate of data is improved.

Furthermore, by further compressing the data of the first block and the compressed information of the second and subsequent blocks using a specific compression means, higher compression becomes possible.

Furthermore, when restoring data in the forward and reverse directions, the compressed multiple data blocks are combined until they reach a specific capacity, and that data group is used as a new data block. Compared to a method in which blocks are separated by a non-data area IBG, the non-data area can be reduced by using information that cannot exist in compressed data (inter-block identification mark), which further improves the efficiency of the recording medium. be able to. Further, by restoring the data of the first block of the compressed data group in advance and preparing it as a basis for restoring the second and subsequent blocks, it becomes possible to carry out the restoration process quickly.

The compression device and the decompression device described above can be used by being incorporated into the MT controller interposed between the computer system central processing unit and the magnetic tape recording device or the control part of the magnetic tape recording device. It can be used by being incorporated into other recording devices, recording device controllers, communication devices, etc.

[Effects of the Invention] As described above, the present invention provides a data compression method, a data compression device, and a data restoration device that can further improve the data compression rate and improve the usage efficiency of recording media, etc. can be provided.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the configuration of a data compression device according to the embodiment, FIG. 2 is a block diagram showing the configuration of the data decompression device according to the embodiment, and FIG. 3 is a block diagram showing the configuration of the data compression device according to the embodiment. FIG. 4 is an explanatory diagram showing the relationship between a data string before inter-block compression and a data string subjected to inter-block compression. FIG. 4 is an explanatory diagram showing a state of recording on a medium using a conventional data compression technique. 3... Buffer memory, 6... Repeated character compression,
8... Buffer memory, 9... Comparator. 10...
Control circuit, 13... Same data counter, 14... Control bit generator, 43... Buffer memory, 44...
- Restoration circuit, 45... Buffer memory, 46... Control circuit, 49... Restoration circuit.

Claims (1)

[Claims] 1. A compression method for a plurality of data blocks, in which a data string constituting a data block other than a predetermined data block is sequentially compressed into data matching a data string constituting the predetermined data block. A data compression method characterized in that: is represented by a data string replaced with information indicating the number of consecutively matched data, and identification information indicating the position of the information of the number of data in the replaced data string. 2. means for comparing each data string of a predetermined data block with a data string of a data block other than the predetermined data block in a predetermined data group consisting of a plurality of data blocks; a counting means for counting the number of consecutive identical data in the same case;
means for creating a new data string by replacing continuous identical data with the count value counted by the counting means; and setting a control bit corresponding to the count value position in the new data string in a predetermined control information area. means for creating a compressed data block corresponding to a data block other than the predetermined data block from the new data string and control information that is information in the control information area; and a compressed data block and the predetermined data block. and means for creating a compressed data group. 3. A count value that is the number of consecutive data blocks other than the predetermined data block in a data group consisting of a plurality of data blocks that match the data string constituting the predetermined data block. a compressed data block consisting of a data string replaced by , and control information in which a control bit corresponding to the position of the count value in the replaced data string is set; and the predetermined data block. An attempt is made to restore the count value in the replaced data string of the data string of the predetermined data block of the same data group by referring to the control bit of control information of the compressed data block in the group. From the position corresponding to the data position of the data block,
A data restoration device characterized in that it has means for replacing data with the data extracted for the count value.