JP2007213198A - Table generation method for data retrieval - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a method for generating a table in which the ID and positions of files are stored with an interval close to the optimal storage interval without preliminarily checking the total number of files recorded in a storage medium. <P>SOLUTION: In a process for recursively retrieving a directory in a storage medium, and for listing files, IDs are successively applied to the listed files, and the set of the IDs and the files is stored in a table. When the table has reached predetermined length, the existing elements of the table whose IDs do not include the multiple of 2 are overwritten by the listed files whose IDs include the multiple of 2, so that it is possible to prevent the table from being made longer any more. When all the elements of the table have IDs including the multiple of 2, those having IDs not including the multiple of 4 are overwritten. Afterwards, the similar operation is repeated until the listing of the files is completed. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a system that assigns an ID to data sequentially supplied to the system and handles the data using the ID. More specifically, the present invention relates to a technique for generating and speeding up a table that stores a plurality of IDs associated with recording positions of data indicated by the IDs.

  Currently, a DCF standard (Design rule for Camera File system) for recording an image taken with a digital camera on a recording medium, or a printing device (for example, a printer) and an image recording device (for example, a digital camera) are connected to There is a PictBridge standard for specifying and printing an image recorded in an image recording apparatus. In either standard, image files are generally stored on a recording medium formatted with the FAT file system. However, when the number of files increases in the FAT file system, it takes time to access the file that is entered later. There is a problem, and several techniques have been devised to solve the problem.

  First, a unique medium ID that is not duplicated in the image recording apparatus (in the recording medium) is assigned to the directory name and the file name, a table is created by associating the media unique ID with the recording position, and stored in a predetermined position. The target file search is speeded up (Patent Document 1).

  The PictBridge standard uses the fact that the ID for identifying a file between the printing apparatus and the image recording apparatus is a 32-bit width and that the file entry value of the image file on the recording medium is the same 32-bit width. By specifying the file entry value as an ID as it is, high-speed file access is realized.

The third is a method for assigning IDs of files and therefore IDs when searching directories recursively. (Patent Document 2 and Patent Document 3)
JP 2004-302853 A JP 2004-295182 A JP-A-9-139915

  Each of the above three ID generation methods has the following problems. The problem with the first method is that the address of a file entry cannot be expressed in 32 bits when the storage medium has a large capacity. The problem with the second method is that only files that are predetermined such as the DCF standard can be communicated. The problem with the third method is that it takes a long time to reach a file located behind the file in the case of recursively searching the directory.

  Here, as one of the methods for solving the problem of the third method, there is a method of improving the file search speed using a cache of the address of the directory area (Patent Document 2). This method certainly improves the file search speed, but it does not solve the problem that it takes more time to get to the back file.

  As another method for solving the problem of the third method, there is a method using a table storing a plurality of sets of file positions and IDs. A technique using a table is disclosed as a technique for searching for a frame at an arbitrary position in a compressed video stream (Patent Document 3), and this technique can be applied to this problem.

  However, when generating a table that stores a plurality of pairs of file positions and IDs, if a large number of files are recorded in the storage medium, it is difficult to store the positions of all the files in the table. It is necessary to select files to be stored and files not to be stored.

  Here, in order to be able to quickly reach a file located behind, that is, a file with a large ID number, it is necessary to store the ID of the file storing the position in the table at equal intervals so as not to be biased. There is. For example, when generating a table storing eight file positions for a storage medium in which 800 files are recorded, eight files with ID = 0, 100, 200, 300, 400, 500, 600, 700 are stored. Ideally the location is stored.

  Here, in order to select files by the above-described method, it is necessary to obtain the total number of files recorded on the storage medium. However, if there are a large number of directories and files on the storage medium, it takes a lot of time. There are challenges.

A similar problem occurs when generating a table that stores a certain number of positions and frame numbers of each frame in the moving image stream in parallel with the recording of the compressed moving image stream. That is, since the user who is recording the moving image does not know when the recording ends, it is unknown how many frames the compressed moving image stream is finally formed. Therefore, it is not possible to determine which frame position is stored in the table.

In order to solve the conventional problem, a table generation method for data search according to claim 1 of the present invention is sequentially supplied with ID assigning means for assigning IDs to data positions on a recording medium. Table storage position determining means for storing the data position and the ID as a table in a buffer memory in the order of data, the table storage position determining means after the storage in the table has been completed, sampling interval N, The position to be stored in the table is determined based on the value of the ID to be newly stored.

  Furthermore, in the table generation method for data retrieval according to claim 2 of the present invention, the sampling interval N starts from 1, and every time storage in the table is completed, the sampling interval N is doubled. It is a feature.

  Furthermore, in the table generation method for data retrieval according to claim 3 of the present invention, the table storage position determining means is configured so that the newly assigned ID is already a multiple of the sampling interval N. An ID value table that is not a multiple of the sampling interval N stored in the table is overwritten, and if the newly assigned ID is not a multiple of the sampling interval N, it is not stored in the table. It is what.

Furthermore, in the table generation method for data retrieval according to claim 4 of the present invention, the data is a file in a storage medium attached to the system,
The ID is the number of other files that exist until a directory is recursively searched to reach the file.

  Furthermore, in the table generation method for data retrieval according to claim 5 of the present invention, the data is one sample of a moving picture or audio stream supplied to the system, and the ID is from the head of the stream to the sample. It is characterized by the elapsed time.

  Furthermore, a table generation method for data search according to claim 6 of the present invention is characterized in that the position of a file in a storage medium is expressed by a file path and stored in the table.

  Further, according to a seventh aspect of the present invention, there is provided a table generation method for data retrieval, wherein a file position in a storage medium is expressed by a file entry address and a head address of a directory area for storing the file and stored in the table. It is characterized by doing.

  Furthermore, in the table generation method for data retrieval according to claim 8 of the present invention, the data is a GOP of an MPEG video stream compressed by MPEG which is an international standard of a moving picture coding system, and the ID and Is the number of GOPs existing in the section from the beginning of the stream to the GOP.

  Furthermore, the table generation method for data retrieval according to claim 9 of the present invention is characterized in that the files in the storage medium are managed by a FAT file system based on the JIS-X-0605 standard. is there.

  According to the table storage position determination means for data retrieval of the present invention, the file ID and position are stored at an interval close to the optimal storage interval without checking the total number of files recorded in the storage medium in advance. To create a table.

  Further, according to the table storage position determining means for data retrieval of the present invention, the position and frame number of each frame in the moving picture stream are fixed in parallel with the recording of the compressed moving picture stream at intervals close to the optimum storage interval. A table for storing numbers can be generated.

  Hereinafter, an embodiment of a table storage position determining means for data search according to the present invention will be described in detail with reference to the drawings.

  FIG. 1 shows a system configuration of a table storage position determination means for data search in the first embodiment of the present invention.

  In FIG. 1, directories and files are recorded on the storage medium 101. The file enumeration means 102 enumerates files recorded in the storage medium 101 by recursively searching directories. The file ID assigning unit 103 assigns a unique ID to the file found by the file listing unit 102. The assigning method assigns IDs in the order in which the file enumeration means finds them, that is, assigns ID = 0 to the first found file and assigns ID = 1 to the next found file. The table storage position determination means 104 determines which element of the table prepared in the buffer memory 105 is to store the sequentially supplied file path and ID pair. Here, it is assumed that the buffer memory 105 can store a table having a combination of a file path and a file ID as one element up to a length of 8.

  FIG. 2 shows directories and files recorded in the storage medium 101 in this embodiment, and also shows IDs assigned to the respective files. Here, an ID is assigned to each file in the order of the files when the directory is recursively searched, that is, in the order in which the file enumeration means 102 finds each file.

FIG. 3 shows the contents of the table when enumerating files with ID = 0-7. Since the buffer memory 105 in this embodiment can store a table up to a length of 8, all files with ID = 0 to 7 are stored in the table.

  FIG. 4 shows the contents of the table when the files of ID = 8 to 15 are listed in addition to the state of FIG. Since there is no storage area for further storing 8 files with ID = 8 to 15 from the state of FIG. 3, the table generation operation is changed as follows. That is, the existing element (ID = 0 to 7) of the table is overwritten with the one whose ID is not a multiple of 2 with the newly supplied file (ID = 8 to 15) whose ID is a multiple of 2. . Thereby, the table of FIG. 4 is generated.

  FIG. 5 shows the contents of the table when the files of ID = 16 to 31 are listed in addition to the state of FIG. In the state of FIG. 4, since there is no table element whose ID is not a multiple of 2, a newly supplied file cannot be stored in the table. Therefore, the table generation operation is changed so that existing elements of the table whose ID is not a multiple of 4 are overwritten with newly supplied files whose ID is a multiple of 4. Thereby, the table of FIG. 5 is generated.

  6 overwrites the same operation from the state of FIG. 5, that is, the existing element of the table whose ID is not a multiple of N with the newly supplied file whose ID is a multiple of N. When there are no existing elements in the table whose ID is not a multiple of N, the contents of the table are shown when N is doubled and all files are listed.

  As described above, in the first embodiment, the existing elements of the table are overwritten with those whose ID is not a multiple of N with the newly supplied file whose ID is a multiple of N, and at this time, the table If there is no existing element of which ID is not a multiple of N, N is doubled so that the optimal storage interval can be obtained without checking the total number of files in the storage medium in advance. A table for storing files at intervals close to can be obtained.

  Finally, a procedure for searching for a file with ID = 250 will be described using the table of FIG. First, the element closest to 250 out of IDs smaller than 250 is searched from the table. Since this corresponds to the element with ID = 224, the file with ID = 250 is searched using the file indicated by ID = 224, that is, “/ZZZ/74.JPG” as the search start position.

  During the search, an ID is added each time a file is found, and the file found when ID = 250 is the target file, but the end of the current directory is reached with the file with ID = 249. Therefore, move to the parent directory. Here, the location of the parent directory area is immediately determined from the file path stored in the table. That is, since the current directory is “/ ZZZ /”, its parent directory is “/”. Thereby, it is possible to move to the head 201 of the parent directory area.

  However, in order to continue the search, it is necessary to move to the position 202 where the entry of the current directory is recorded in the parent directory area instead of moving to the top 201 of the parent directory area. Here, the position 202 where the entry of the current directory is recorded is also found from the file path. That is, since the name of the current directory is “ZZZ”, it is only necessary to search for entries whose name is “ZZZ” among the entries recorded in the parent directory area.

  When the move is complete, the search is resumed. Here, since the current ID is 249, the first file found after the search is resumed is the file with ID = 250. This is "/50.JPG".

  By storing the file path in the table in this way, the directory can be recursively searched using the file stored in the table as the search start position.

  FIG. 7 shows the system configuration of the table storage position determining means for data search in the second embodiment of the present invention.

  In FIG. 7, a storage medium 701 records directories and files in accordance with a FAT file system based on the JIS-X-0605 standard. The file listing unit 702 lists the files recorded on the storage medium 701 by recursively searching the directory. The file ID assigning unit 703 assigns a unique ID to the file found by the file listing unit 702. The assigning method assigns IDs in the order in which the file enumeration means finds them, that is, assigns ID = 0 to the first found file and assigns ID = 1 to the next found file. The table storage position determination unit 704 determines which element of the table prepared in the buffer memory 705 stores the set of the file entry address that is sequentially supplied, the head address of the directory area that stores the file entry, and the ID. decide. Here, it is assumed that the buffer memory 705 can store a table up to a length of 8.

  FIG. 8 shows directories and files recorded on the storage medium 701 in this embodiment, and also shows IDs assigned to the respective files. Here, an ID is assigned to each file in the order of the files when the directory is recursively searched, that is, in the order in which the file enumeration means 702 finds each file. Further, it is assumed that the root directory area 801 starts from address 0, and the “/ ZZZ /” directory area 802 starts from address 00001000 (hexadecimal number).

  FIG. 9 shows a table generated in the buffer memory 705 when all the files shown in FIG. 8 are listed. The table generation procedure is the same as that described in the first embodiment of the present invention. The difference from the first embodiment is that the file path is stored in the table in the first embodiment, but in the second embodiment, the file entry address and the head address of the directory area storing the file entry are changed. It is a point stored in a table. In the JIS-X-0605 standard, the file entry address in FIG. 9 is determined by using 32 bytes per entry.

  Here, a procedure for searching for a file with ID = 250 using the table shown in FIG. 9 will be described. As in the first embodiment, first, among the elements whose ID is smaller than 250, the closest element to 250 is searched from the table. Since this corresponds to the element with ID = 224, the file with ID = 250 is searched using the entry address of the file indicated by ID = 224, that is, 000019A0 (hexadecimal number) as the search start position.

  During the search, ID is incremented by 1 every time a file is found, and the file found when ID = 250 is the target file, but the end of the current directory is reached with the file with ID = 249. So move to the parent directory. Here, it is necessary to investigate where the parent directory area is. However, in the FAT file system based on the JIS-X-0605 standard, an entry 804 for storing the address of the parent directory area of the directory is provided near the head of the directory area. Exists. In addition, the head address of the current directory area 802 is stored in the table. By these, it is possible to move to the head of the parent directory area 801.

  However, in order to continue the search, it is necessary not to move to the top of the parent directory area 801 but to move to the position 803 where the entry of the current directory is recorded in the parent directory area. Here, the position 803 where the current directory entry is recorded is also determined from the head address of the directory area 802. That is, the directory entry recorded in the parent directory area may be searched for using the directory area 802.

  When the move is complete, the search is resumed. Here, since the current ID is 249, the first file found after the search is resumed is the file with ID = 250. This is "/50.JPG".

  By storing the file entry address and the head address of the directory area that stores the file entry in the table in this way, the directory can be recursively searched using the file stored in the table as the search start position, and the file path can be changed. Compared with the storing method, the table size can be saved.

  FIG. 10 shows the system configuration of the table storage position determining means for data search in the third embodiment of the present invention.

  In FIG. 10, the imaging unit 1001 sequentially supplies captured video to the encoding unit 1002, and the encoding unit 1002 encodes the sequentially supplied captured video into MPEG format stream data. Here, the stream data body of the encoded data is recorded on the storage medium 1004 by the storage medium writing means 1003, and the address of each GOP is supplied to the GOP number assigning means 1005. The GOP number assigning unit 1005 sequentially assigns numbers to the GOP addresses that are sequentially supplied from the 0th order, and sequentially supplies pairs of GOP addresses and GOP numbers to the table storage position determining unit 1006. The table storage position determination unit 1006 stores a set of sequentially supplied GOP addresses and GOP numbers in the buffer memory 1007. Here, it is assumed that the buffer memory 1007 can store a table that can store up to eight sets of GOP addresses and GOP numbers.

  FIG. 11 shows a table when the GOP addresses with GOP numbers 0 to 7 are supplied to the table storage position determining means. Here, the GOP address is expressed by the number of offset bytes from the beginning of the stream to the GOP. Further, since the buffer memory 1007 can store up to eight sets of GOP numbers and GOP addresses, all the supplied GOP numbers and GOP addresses are stored in a table.

  FIG. 12 shows the contents of the table when GOP addresses with GOP numbers = 8 to 15 are supplied in addition to the state of FIG. Since the table in FIG. 11 uses up the buffer memory 1007, the newly supplied GOP address cannot be stored as it is, so the table generation operation is changed as follows. That is, the existing elements (GOP number = 0 to 7) of the table whose GOP number is not a multiple of 2 and the GOP number of the newly supplied GOP address (GOP number = 8 to 15) is a multiple of 2. Overwrite with things. Thereby, the table of FIG. 12 is generated.

  FIG. 13 shows the contents of the table when GOP addresses with GOP numbers = 16 to 31 are supplied in addition to the state of FIG. Since there is no element in the table of FIG. 12 whose GOP number is not a multiple of 2, a newly supplied GOP address cannot be stored in the table. Therefore, the table generation operation is changed such that existing elements of the table whose GOP number is not a multiple of 4 are overwritten with newly supplied GOP addresses whose GOP number is a multiple of 4. Thereby, the table of FIG. 13 is generated.

  FIG. 14 overwrites the same operation from the state of FIG. 13, that is, the existing element of the table whose GOP number is not a multiple of N with the newly supplied GOP address whose GOP number is a multiple of N. At this time, if there are no existing elements in the table whose GOP number is not a multiple of N, N is doubled, and the contents of the table when the GOP addresses up to GOP number 999 are supplied are shown. Show.

  As described above, in the third embodiment, the existing elements of the table are overwritten with the GOP number that is not a multiple of N among the newly supplied GOP addresses, and the GOP number is a multiple of N. At this time, if there is no GOP number that is not a multiple of N among the existing elements of the table, N is doubled so that the number of GOP addresses supplied is not known in advance by performing table generation. A table for storing GOP addresses at intervals close to the optimal storage interval can be obtained.

  Note that the procedure for searching for a frame at an arbitrary point in an MPEG video stream using the table shown in the third embodiment is known from Patent Document 3 and is not the main point of the present invention. Omitted.

  The table storage position determination means for data search according to the present invention has the effect of generating a table for storing data positions at intervals close to the optimal storage interval without examining the total length of the search target data group in advance, This is useful for a system for identifying each file in a storage medium by ID, a system for searching a frame at an arbitrary point of a compressed moving image, and the like.

The system block diagram of the table storage position determination means in Example 1 of this invention The file tree figure of the storage medium of the table storage position determination means in Example 1 of this invention The table figure which stores the file to ID = 7 of the table storage position determination means in Example 1 of this invention. The table figure which stores the file to ID = 15 of the table storage position determination means in Example 1 of this invention. The table figure which stores the file to ID = 31 of the table storage position determination means in Example 1 of this invention. The table figure which stores all the files of the table storage position determination means in Example 1 of this invention The system block diagram of the table storage position determination means in Example 2 of this invention FIG. 7 is a file tree diagram of a storage medium of a table storage position determination unit in Embodiment 2 of the present invention. The table figure which stores all the files of the table storage position determination means in Example 2 of this invention The system block diagram of the table storage position determination means in Example 3 of this invention The table figure which stores the GOP address to GOP number 7 of the table storage position determination means in Example 3 of this invention. The table figure which stores the GOP address to GOP number 15 of the table storage position determination means in Example 3 of this invention. The table figure which stores the GOP address to GOP number 31 of the table storage position determination means in Example 3 of this invention. The table figure which stores the GOP address to GOP number 999 of the table storage position determination means in Example 3 of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 101 Storage medium in Example 1 102 File enumeration means in Example 1 103 File ID allocation means in Example 1 104 Table storage position determination means in Example 1 105 Buffer memory in Example 1 201 Root directory area in Example 1 202 Directory entry for storing the file of ID = 224 in the second embodiment 701 Storage medium in the second embodiment 702 File listing means in the second embodiment 703 File ID assigning means in the second embodiment 704 Table storage location determining means in the second embodiment 705 Buffer memory in the second embodiment 801 Root directory area in the second embodiment 802 Directory area for storing the file of ID = 224 in the second embodiment 803 In the second embodiment Entry of directory storing file with ID = 224 804 Entry for storing address of parent directory area of directory storing file with ID = 224 in embodiment 2 1001 Imaging means in embodiment 3 1002 Encoding means in embodiment 3 1003 Storage medium writing means in embodiment 3 1004 Storage medium in embodiment 3 1005 GOP number assigning means in embodiment 3 1006 Table storage position determining means in embodiment 3 1007 Buffer memory in embodiment 3

Claims (9)

ID assigning means for assigning ID to the data position on the recording medium;
Table storage position determining means for storing the data position and the ID as a table in a buffer memory in the order of sequentially supplied data;
The table storage position determining means determines a position to be stored in the table based on a sampling interval N and a value of the ID to be newly stored after the storage in the table is completed. 2. The table generation method according to claim 1, wherein the sampling interval N starts from 1 and the sampling interval N is doubled every time storage in the table is completed. When the newly assigned ID is a multiple of the sampling interval N, the table storage position determination means overwrites a table of ID values that are not already a multiple of the sampling interval N and are already stored in the table. 2. The table generation method according to claim 1, wherein if the newly assigned ID is not a multiple of the sampling interval N, the ID is not stored in the table. The data is a file in a storage medium attached to the system,
2. The table generation method according to claim 1, wherein the ID is the number of other files existing until a directory is recursively searched to reach the file. The data is a sample of a video or audio stream supplied to the system,
The table generation method according to claim 1, wherein the ID is an elapsed time from the head of the stream to the sample. 5. The table generation method according to claim 4, wherein the position of the file in the storage medium is expressed in a file path and stored in the table. 5. The table generation method according to claim 4, wherein the position of the file in the storage medium is stored in the table by representing the address of the file entry and the head address of the directory area in which the file is stored. The data is a GOP of an MPEG video stream compressed by MPEG, which is an international standard for moving picture coding.
6. The table generation method according to claim 5, wherein the ID is the number of GOPs existing in a section from the head of the stream to the GOP. 8. The table generation method according to claim 7, wherein files in the storage medium are managed by a FAT type file system based on the JIS-X-0605 standard.

Images