Ronald Fagin. On an authorization mechanism. ACM Transactions on Database Systems, 3(3):310--319, September 1978.  Home/Search   Document Not in Database   Summary   Related Articles   Check

.... stops retrieving new objects when it finds a threshold grade G such that (1) at least c objects with grade G or higher have been identified, and (2) no unretrieved object can have a 5The problem of sequencing the order of accesses to subfiles of transposed files is also related in a similar way [2]. 32 grade greater than G. Unlike our Rank algorithm, TA does not determine G using selectivity statistics, but rather refines G s value dynamically based on the grades of retrieved objects. As a result, TA never needs to restart a query. Other recent work includes Bruno et al. s Upper ....

D. S. Batory. On searching transposed files. ACM Transactions on Database Systems, 4(4), Dec. 1979. 33

....pairwise consistency. In CSP terms, pairwise consistency requires that a pair of constraints be compatible. More formally, Definition6. Beeri et al. 1] A CSP (X; D;V; S) is pairwise consistent if for any V i ; V j 2 V , Pi V i (S i 1 S j ) S i and Pi V j (S i 1 S j ) S j . Gyssens [7] generalizes pairwise consistency into k wise consistency (also called inter k consistency) Pairwise consistency is then identical to 2 wise consistency. Definition7. Gyssens [7] A CSP (X; D;V; S) is k wise consistent if for any V p1 ; V p2 ; V pk Gamma1 ; V pk 2 V , Pi Vp k (1 k i=1 ....

....is pairwise consistent if for any V i ; V j 2 V , Pi V i (S i 1 S j ) S i and Pi V j (S i 1 S j ) S j . Gyssens [7] generalizes pairwise consistency into k wise consistency (also called inter k consistency) Pairwise consistency is then identical to 2 wise consistency. Definition7. Gyssens [7]) A CSP (X; D;V; S) is k wise consistent if for any V p1 ; V p2 ; V pk Gamma1 ; V pk 2 V , Pi Vp k (1 k i=1 S p i ) S pk . In words, a set of k constraints is k wise consistent if any tuple in any constraint has a consistent extension to all the variables involved in these k ....

M. Gyssens. On the complexity of join dependencies. ACM Transactions on Database Systems, 11(1):81--108, 1986.

....levels to manage storage of differing speed to provide good performance across the entire spectrum of resolutions. The quadtree representation of raster images is an inherently multi resolution data structure. Structures such as a bit striped transposed representation of sets of ordered values [32, 33] may allow multi resolution to be applied to common, built in types more efficiently. These issues, however, are beyond the scope of this paper. ....

D. S. Batory. On searching transposed files. ACM Transactions on Database Systems, 4(4):531--544, December 1979.

....extension are stored in another. The first entry in each of the files corresponds to the first pair of objects, the second entry to the second pair, and so on. There is no need for any additional stored information This vertically fragmented data structure has been termed a transposed file [Bat79] 3 Adapting Jive Join to Object Oriented Databases Jive Join is an algorithm for performing joins in a relational database system using a join index [RL95b] In this section we adapt the algorithm to apply to object oriented databases. We will measure the cost of the new algorithm using the ....

D. S. Batory. On searching transposed files. ACM Transactions on Database Systems, 4(4):531--544, 1979.

....the database to be treated like a relational database, thus harnessing known and standard access methods and operations for record manipulation. We are now working on extending the compression techniques to relational databases in general. 6. 2 Attribute Transposition Attribute transposition [4, 27] stores a relation as a collection of contiguous attribute columns, where all values for an attribute domain are stored together. Work in this area has concentrated on different schemes for encoding attributes columns. Since attribute values are repeated, the standard coding methods may be used to ....

D. S. Batory. On Searching Transposed Files. ACM Transactions on Database Systems, Vol. 4, No. 4, pp. 531--544, December 1979.

....the search costs. On the other hand, the problem of determining an optimal set of conditions to search arises naturally when optimizing single table queries with multiple indexes [RR82, Moh90] The problem of sequencing the order of accesses to subfiles of transposed files is also closely related [Bat79] However, in the above contexts, the probing cost is either zero or is independent of the predicates. Our approach to defining the execution space is similar in spirit to [Moh90] but our problem is more complex since probe costs can be significant as well as varied. When the filter condition is ....

Don S. Batory. On searching transposed files. ACM Transactions on Database Systems, 4(4), December 1979.

....efficient performance is the use of a vertically partitioned data structure for the join result. We do not assume that the inputs are vertically partitioned. Attributes from the first input relation are stored in a separate file from those of the second input relation, using transposed files [2]. Attributes that are common are placed arbitrarily in one of the two vertical fragments. There is a oneto one correspondence between records in each vertical partition: The nth record in the first vertical fragment matches the nth record in the second. We will argue that such a representation ....

....disk seeks between accesses. The input relations may reside on the same disk. This kind of configuration is recommended by most commercial vendors. 2. 1 A Vertically Partitioned Data Structure for the Join Result We use a vertically partitioned data structure known as a transposed file [2] to store the join result. Attributes from R 1 that are present in the join result are stored in a separate file (denoted JR 1 ) from those of R 2 (which are in JR 2 ) In Section 8, where we consider joins of three or more relations, there will be JR 3 , JR 4 , etc. Join attributes that are ....

Batory, D. S. On searching transposed files. ACM Transactions on Database Systems 4, 4 (1979), 531--544.

....a smallMap shown in figures 3 and 4. The small map representation is attempting space efficiency and sacrificing cpu efficiency. The entries in the map are held in the map order, and accessed via a binary search algorithm. The map is held in vectors, one per column (cf. Batory s transposed files [Bat79] Consequently, all the elements in a vector are of the same type, and hence this fits with Napier88 s type rules. This choice is also motivated by avoiding repeated object overheads in the store, and avoiding unnecessary transfers to store, as most of a search will be localised to one vector. ....

D.S. Batory. On searching transposed files. ACM Transactions on Database Systems, 4:531--544, December 1979.

....levels to manage storage of differing speed to provide good performance across the entire spectrum of resolutions. The quad tree representation of raster images is an inherently multi resolution data structure. Structures such as a bit striped transposed representation of sets of ordered values [31, 32] may allow multi resolution to be applied to common, built in types more efficiently. Acknowledgements Comments by the referees and the UT database research group have greatly improved this paper. ....

D. S. Batory. On searching transposed files. ACM Transactions on Database Systems, 4(4):531--544, December 1979.

....the search costs. On the other hand, the problem of determining an optimal set of conditions to search arises naturally when optimizing single table queries with multiple indexes [20, 21] The problem of sequencing the order of accesses to subfiles of transposed files is also closely related [22]. However, in the above contexts, the probing cost is either zero or is independent of the predicates. Our approach to defining the execution space is similar in spirit to [21] but our problem is more complex since probe costs can be significant as well as varied. When the filter condition is ....

Don S. Batory. On searching transposed files. ACM Transactions on Database Systems, 4(4), December 1979.

....high. Like Jive join, Stripe join uses a vertically partitioned data structure for the join result. We do not assume that the inputs are vertically partitioned. Attributes from the first input relation are stored in a separate file from those of the second input relation, using transposed files [1]. Attributes that are common are placed arbitrarily in one of the two vertical fragments. There is a one to one correspondence between records in each vertical partition: The nth record in the first vertical fragment matches the nth record in the second. Such a representation has a negligible ....

....vendors. 2 Actually, for the Stripe join algorithm one can store the join index on the same disk as the input relations with no loss of performance. 2. 1 A Vertically Partitioned Data Structure for the Join Result We use a vertically partitioned data structure known as a transposed file [1] to store the join result. Attributes from each R i that are present in the join result are stored in a separate file (denoted JR i ) Join attributes that are common to more than one relation are placed arbitrarily in one of the vertical fragments. The first entry in each of the files corresponds ....

D. S. Batory. On searching transposed files. ACM Transactions on Database Systems, 4(4):531--544, 1979.

No context found.

On Mismatches Between Ontologies KRAFT Paper KP10, PRSV 2.4 / Oct96, Page 20 Subrahmanian, V.S., (1994). Amalgating Knowledge Bases, ACM Transactions on Database Systems, Vol.19, No. 2, pp.291-331.

No context found.

Ronald Fagin. On an authorization mechanism. ACM Transactions on Database Systems, 3(3):310--319, September 1978.

No context found.

R. Fagin. On an Authorization Mechanism. ACM Transactions on Database Systems, 3(3):310--319, 1978.

No context found.

Don S. Batory. On Searching Transposed Files. ACM Transactions on Database Systems, 4(4):531--544, 1979.

No context found.

R. Fagin. On an Authorization Mechanism. ACM Transactions on Database Systems, 3(3):310--319, 1978.

No context found.

M. Gyssens, On the Complexity of Join Dependencies, ACM Transactions on Database Systems 11(1): 81--108, 1986.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC