| 摘要 |
<p>A method of sorting a database comprising a plurality of records and a plurality of data items, in which each record comprises an identifier, first and second data variables and paths pointing to data items being the value of said data variables, in which the database further comprises a first frequency for each path of the first data variable, and a second frequency for each path of the second data variable, comprising the following steps:
- creating an intermediate array comprising a section for each value of the second data variable, which sections comprise identifier storage locations equal in number to the corresponding second frequency;
- storing the identifier of each record into the section of the intermediate array corresponding to the value of its second data variable;
- creating a final array comprising sections for each value of the first data variable, which sections comprise identifier storage locations equal in number to the corresponding first frequency;
- storing the identifier of the records appearing in a first section of the intermediate array into the section of the final array corresponding to the value of its first data variable;
- identifying break points in the final array corresponding to the last record stored into each section thereof in the previous step;
- repeating the previous two steps for each further section of the intermediate array.
- creating a break point index comprising the break points identified in each repetition.</p> |
