JPS63136222A - Data alignment processing system fitted with variable length label - Google Patents

Abstract

PURPOSE:To minimize a memory area and the comparison/transfer data value and to shorten the processing time by aligning the corresponding data based on the relevant original labels registered in a table after the labels larger than the prescribed length are compressed. CONSTITUTION:When a data input part 1 supplied the unaligned data including the variable length labels to a label length fixing part 2 via a file 7, the corresponding relation between the variable length labels and the original labels is registered into a table table 3 with these variable length compressed to the fixed length in case the variable length labels are larger than the prescribed length. While the blank codes are added when the variable labels are smaller than the prescribed value and these variable lengths are turned into the fixed value. When this length fixing process is through, a fixed length label data aligning part 4 is started and the arrangement processing is repeated for two data containing the labels of the compressed and expanded fixed lengths according to the difference of label lengths. Then a label restoring part 5 is started and the original variable length labels are restored in response to the compressed or non-compressed state of the fixed length labels. Then the arranged data containing the variable length labels undergone no length fixing process via expansion/ compression are applied to the corresponding file 8 via a data output means 6.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of the Invention The present invention relates to a method for sorting variable-length labeled data used for language processing and utilities in information processing systems.

In conventional technical information processing systems, it is often necessary to perform a data sorting process in which data to which variable-length names are attached is rearranged in a predetermined order using variable-length names as keys.

Conventionally, in such data sorting processing methods, each name tag of variable length was compared with a fixed length according to the number of digits of the maximum length, or it was compared with a variable length while keeping the length of the name tag in mind. , only the name tag part is registered in a table and the table is referenced every time a comparison is made.

Problems to be Solved by the Invention Among the conventional data sorting processing methods mentioned above, the method in which name tags are set to a fixed length according to the number of digits of the maximum length and compared is a method that adjusts the area of all name tags to the maximum length There is a problem in that not only a large storage area is required, but also the amount of data to be transferred and compared increases.

In addition, in the method of comparing variable lengths while keeping in mind the length of the name tag, it is difficult to increase the comparison speed because the comparison target is variable length, and the amount of data transferred is also variable length, making operations complicated. There is a problem.

Furthermore, with the method of registering name names in a table and referring to the table each time a comparison is made, the comparison speed decreases due to the amount of time required to access the table, and a large amount of memory is required to register all name names in the table. There is a problem that.

Means for Solving the Problems The method of sorting data with variable-length name tags of the present invention is that for name tags attached to each data that are longer than a predetermined length, the correspondence with the original name tag is stored in a name table. A name fixed length conversion unit compresses the name into a fixed length while registering it, and adds a predetermined code to the name less than a predetermined length to expand it to the predetermined length. Compare the name marks, and if they match, if one or both are compressed, the corresponding original name name registered in the name name table is used, and in other cases, the corresponding name name is compared based on the fixed-length name name. A fixed-length name-tagged data arrangement unit that performs data sorting processing, and a fixed-length name tag attached to each data sorted in a predetermined order by this fixed-length name-tagged data name arrangement unit. By providing a name name restoration means that restores the name name name to a variable-length original name name while referring to the name name table, the step of comparing name names, which takes the most time and requires a large amount of hardware, can be performed based on compressed fixed-length name names. configured to run.

Embodiment FIG. 1 is a conceptual diagram of a method according to an embodiment of the present invention, in which 1 is a data input section, 2 is a fixed-length name labeling section, 3 is a name table, and 4 is a data arrangement with fixed-length name tags. Department, 5 is nameplate restoration department,
6 is a data output section, 7 is a storage file for unsorted data,
8 is a storage file for sorted data.

The data input unit 1 inputs unsorted data with variable-length name tags attached from the storage file 7 for unsorted data, and passes it to the name tag fixed length conversion unit 2 . The fixed name name length conversion unit 2 checks the length of the variable length name tag attached to the input unsorted data, and if it is longer than a predetermined length, the name name fixed length conversion unit 2 stores the correspondence with the original name name in the name name table. Compress to fixed length while registering. Furthermore, if the length of the variable-length name tag attached to the unsorted data is less than a predetermined value, the name fixed length generator 2 expands it to the above-mentioned predetermined length by adding a predetermined code such as a blank space. By doing this, the length is fixed.

FIG. 2 is a conceptual diagram showing an example of a fixed-length name name created by the name name fixed-length conversion unit 2 and the contents registered in the name name table 3.

Each landmark is compressed to a fixed length of N bits. Of these N bits, the first n1 bits are a valid comparison part, and the latter n2 bits are a comparison invalid part that stores information about the presence or absence of compression and the correspondence with the original data.

In the case of the original name mark rAJ as illustrated in the top row of Figure 2,
Since its length exceeds the comparison effective part of n, bit width, the excess part 'A2
”, and the n1 bit width comparison effective part “AI”
, a compression display flag F1 indicating that this is compressed, and a pointer P of (nt-1) bits in the name table 3 storing the excess portion "A2".
It is compressed into a fixed-length name of N bits with a 2-bit comparison invalid part added. Then, in the area in the name table 3 specified by the pointer P, the excess portion "At" of the name rAJ is registered.

In the case of the original name mark rBJ as illustrated in the second row, its length matches the comparison effective part of n bits width, so the comparison effective part consisting of the original name mark rBJ itself of n bits,
This name name is converted into an N-bit fixed-length name name consisting of an uncompressed display flag F2 indicating that the name name is not compressed, and an invalid pointer P0 consisting of (nz-1) bits of continuous "O".

In the case of the original name mark rcJ as exemplified in the third row of Figure 2,
Since its length is less than the effective comparison part of n1 focus width, a space is added to the original name mark "C" itself to display the comparison effective part expanded to nl bits and that this name mark is not compressed. Uncompressed display flag F2 and (nz-1
) It is converted into an N-bit fixed-length name mark consisting of an invalid pointer P0 consisting of a series of "0" bits.

When the name tag fixed length conversion section 2 completes the above-described name tag fixed length conversion processing for all the unsorted data received from the data input section 1, it starts the fixed length name tag data arrangement section 4.

The data sorting unit 4 with fixed length name tags examines the size relationship of the name tags attached to the two data having fixed length name tags compressed and expanded as described above, and
The sorting process of rearranging the order of both is repeated according to this size relationship. At this time, the fixed-length name tag data sorting unit 4 compares the comparatively valid portions of the fixed-length name tags attached to both data, and if the two match, the compressed display flag F,
Inspect. When the fixed length name tag data sorting unit 4 detects that one or both of them are compressed from the above inspection results, the fixed length name tag data sorting unit 4 sorts out the size relationship based on the original name tag including the excess portion of the correspondence currently registered in the name tag table 3. In other cases, the data sorting unit 4 with fixed-length name tags sorts both data using only the comparatively valid portions of the fixed-length name tags.

The fixed-length name-tagged data sorting unit 4 activates the name-tag restoring unit 5 after completing the sorting of all the fixed-length name-tagged data by the above-described predetermined sorting.

The name name restoring unit 5 checks whether the fixed length name tags added to each of the data arranged in a predetermined order is compressed based on the compressed display flag F and the non-compressed display flag F2,
If it is compressed, the excess portion is read out from the name table based on the pointer of the comparison invalid portion and combined, thereby restoring the name to the original variable length name. The nameplate restoration unit 5
If spaces are added to the comparison valid part of the uncompressed name tag,
By truncating this, the original variable length name tag is restored.

When the restoration of the variable length name tags described above is completed for all the sorted data, the data output means 6 is activated, and the sorted data with the variable length name tags attached before being converted to a fixed length by expansion/contraction is output to the corresponding file 8. be done.

Effects of the Invention As explained in detail above, the method of sorting data with variable-length name tags of the present invention is that for name tags that are longer than a predetermined length, the correspondence relationship with the original name tag is registered in the name tag table while the data is arranged in a fixed length. Compare the fixed-length names attached to each data, and if both lose once, if one or both are compressed, the corresponding original name is registered in the name table. The data is sorted, and after the sorting process is completed, the fixed-length name tags attached to each data are restored to the variable-length original name tags while referring to the name table above, so the maximum length is fixed. Compared to a method that takes longer, the storage area and the amount of data for comparison and transfer can be minimized, which has the effect of shortening processing time and reducing the amount of hardware.

Furthermore, compared to the method of comparing variable lengths while keeping in mind the length of the name tag, the operation is easier to understand and the processing time can be shortened.

Furthermore, compared to a method in which names are registered in a name table and the table is referred to each time a comparison is made, processing speed and memory capacity can be reduced by the proportion of non-abbreviated names that do not require table reference.

[Brief explanation of the drawing]

FIG. 1 is a conceptual diagram of a method for sorting data with variable-length name tags according to an embodiment of the present invention, and FIG. 2 is a conceptual diagram showing an example of fixed-length processing for name tags. 1...Data input section, 2...Name tag fixed length conversion section, 3.
...Name table, 4...Fixed length name tag data arrangement unit, 5...Name name restoration unit, 6...Data output unit, 7.
・Unsorted data storage file with variable length name label, 8...
- Sorted data storage file with variable length name tags.

Claims (1)

[Claims] In a data sorting method that arranges each piece of data in a predetermined order using a variable-length name tag attached to data as a key, for the name tags attached to each data that are longer than a predetermined length, a fixed name name length converting unit that compresses the correspondence to the original name name in a name name table while compressing it to a predetermined length, and adds a predetermined code to those whose length is less than the predetermined length to expand them to the predetermined length; Compare the fixed-length names attached to each data, and if they match, if one or both are compressed, use the corresponding original name registered in the name table, and in other cases. is attached to each data sorted in a predetermined order by the fixed-length name-tagged data alignment unit that sorts the corresponding data based on the fixed-length name tag. a name name restoring means for restoring a fixed-length name name in a fixed-length name name into a variable-length original name name while referring to the name name table;