ARRANGEMENT FOR DATA COMPRESSION

摘要

System (i) for putting in store an Entity Key (EK) set; (ii) for providing a response to a subsequent encoded input of a particular EK (INPUT SEQUENCE OF OCTET I/O CODE-WORDS REPRESENTING THE EK) - such response being an encoded output of a matching/non-matching condition, a pointer value and/or an EK Sequence number (EKS); and/or (iii) for providing a response to a subsequent encoded input of a particular EKS - such response being an output sequence of I/O code-words representing the corresponding particular EK. The system comprises (i) code conversion means (46) transforming the current EK input into a number of linked digital words (L1 through LZ), which thereby attain individual current link values and jointly a current overall link value representing the EK input in a compact form, compaction being performed using various means (e.g. BIT MAP MEMORY MEANS); (ii) a mandatory first link memory unit (12) holding previously stored EK related first link values (L1); (iii) a non-mandatory intermediate link memory unit (58) holding previously stored EK related intermediate link values (L2 and L3); (iv) a mandatory main link memory unit (14), holding an ordered set of previously stored main link values (L4), each member being related to at least one member of the EK set; and (v) a non-mandatory additional link memory unit (50) holding previously stored EK related additional link values, if any, (L5 through LZ). Use of various data compression techniques, particularly within input code conversion means, facilitates compact storage of large EK sets and contributes to very short response times in search operations, but also to fast sorting at input of additional, previously not stored, Entity Keys. Hence, searching and sorting speeds are reduced, e.g. in data base applications, EKS code-words may also be applied to represent data base file segments per se, e.g. words and/or phrases of a text file, for compact file storage and for fast file transfer.