| 摘要 |
A method whereby a database storage structure is created by selectively applying one or more data compaction methods to fields of a database. A specific compaction method is applied to a field if the field data characteristics satisfy criteria for that compaction method. The compaction methods used are: single-field encoding, where codes are substituted for data values in a field; multiple-field combining, where a single code is substituted for data values from two or more fields; pattern suppression, where recurring character patterns within data values are removed; numeric substitution, where binary values are substituted for numeric character data; and text compression, where codes are substituted for words and phrases in a text field. These compaction methods create compacted records which are reduced storage equivalents of the database records. Translation tables and auxiliary tables created by the compaction methods allow the compacted data to be retranslated into a user-readable format. The compacted records are grouped into storage partitions, each containing compacted records of the same length, to form a database image. The database image advantageously resides in a computer system's mass storage while the translation and auxiliary tables advantageously reside in the computer system's fast access memory. Along with an accessing subsystem, the database image and tables function as a reduced storage equivalent of the original database. |
