摘要 |
PURPOSE:To efficiently compress a text irrespective of an appearance pattern of data, and to softly and exactly restore it. CONSTITUTION:A text stored in a buffer 15 is divided into words, the word contained in an independent work dictionary 11 is converted to its address, outputted successively to an output buffer 9 and the text is compressed. When the comprecessed text is outputted to a disk 14, discriminating information for specifying the dictionary 11 is added by a discriminating information adding part 4. At the time of restoration, whether the dictionary 11 at the time of restoration and compression is the same or not is decided, based on the discriminating information, and in the case it is different, a table 12 for replacing addresses of two dictionaries is prepared. A dictionary address in the text is converted to the corresponding address of the dictionary 11 at the time of restoration by referring to the table 12 by a dictionary address replacing part 7. By inputting the word stored in its address to text data by a sentence restoring part 6, compressed text data is restored. |