G. Graefe and P. O'Neil. Multi-table joins through bitmapped join indices. ACM SIGMOD Record, 24(3):8--11, October 1995.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....product for all store values. An additional index can hold the aggregates for the store dimension. The main disadvantage of this method is that the number of indexes and therefore the space requirements grows exponentially as the number of dimensions increases. Bit mapped indices and variations [OG95, OQ97, WB98] provide another means of handling sparse data. In its simplest form a bit mapped index is a B tree that instead of storing at the leaf pages a list of record ids for each key value, it stores a compressed bit map. There is one such bit map for each value of the key. Each bit map has ....

P. O'Neil and G. Graefe. Multi-Table Joins Through Bitmapped Join Indices. SIGMOD Record, 24(3):8--11, Sept 1995.

....a materialized view MV is defined as cost(MV, QR) N4 v where N4v denotes the number of pages in MV. On the other hand, several types of access methods have been proposed in the literature to support fast access to the fact table in data warehouses [7,20,25] In particular, bitmap join indices [19] on the dimensional attributes are frequently used for efficient joins between a dimension table and the fact table. 21 Considering such index structures, we assume the execution cost of a query block over the fact table to be the number of tuples contained in a given query region over the fact ....

P. O'Neil and G. Graefe, Multi-Table Joins Through Bitmapped Join Indices, SIGMOD Record 24 (3) (1995) 8-11.

....integrity maintenance in the presence of updates is easily enforced. We also show that tuples in the representative instance over U , i.e. tuples over the fact table extended with the relevant information in the constituent dimension tables, can be computed via the snowflake join (cf. star join [OG95]) This implies that the results of queries over a data warehouse which is in SSNF maintain their consistency, i.e. satisfy the induced set of FDs over their schema. We also examine an information theoretic interpretation of the snowflake schema following the work of [Mal86, CP87, Lee87, Mal88] ....

..... n, be in BCNF since the redundancy H(R i X X) in r i # d over R i is expected to be insignificant relative to the potential redundancy in r 0 were it not to be in BCNF. On the other hand, by not decomposing the dimension tables R i query processing can be optimised via the star join [OG95]. The following result is a special case of Theorem 7 in [Mal88] and Theorem 4 in [Lee87] Lemma 4.2 Let R be a snowflake schema and r be a relation over U . Then r = #[R] if and only if H r (U) n # i=0 H(R i ) n # i=1 H(X i ) where R i and X i are as introduced after ....

P.E. O'Neil and G. Graefe. Multi-table joins through bitmapped join indices. ACM SIGMOD Record, 24:8--11, 1995.

....table and the fact table. We need to use Data Definition Language (DDL) to add a column and reload the diagnosis dimension by adding (2 N 1 2 N ) rows and updating the diagnosis key in the fact table; there are no approaches to handle a weighting factor. However, a bitmap index scheme [OG95] can be implemented on each positional attribute, which would improve the query performance in this approach. It is clear that this method would only be applicable when the number of positional attributes is limited and fixed. 3.2.1. Database Sizing for the Positional Flag Attribute Method ....

....the query performance. In order to resolve the problem of LIKE clause, we can enhance the non positional model by incorporating the benefits of positional flag attributes. Additional Boolean attributes can be created for common or frequent diagnoses. See Table 6 for an example. Bitmap indexes [OG95, CI98] can be created for these Boolean attributes to facilitate searching based on these common diagnoses. Unusual or intriguing diagnoses could also be included for specific business intelligence purposes. The hybrid method allows both pattern matching with the LIKE command and an index search through ....

O'Neil, P. and Graefe, G., Multi-table Joins Through Bitmapped Join Indices, SIGMOD Record, Vol. 24, No.3, Sept. 1995, pp. 8-11.

....statistical data, etc. are integrated into a large, consistent data repository, the DWH (data acquisition phase) Specialized modeling concepts such as the multidimensional model, the star and the snowflake schema as well as storage structures like bitmap ( 7] 8] 39] 41] and join indices ([23], 21, are used to store extracted data in the DWH and or in particular data marts (data storage phase) A data mart is a selected part of the data warehouse which contains simple replicas of warehouse partitions or 2 data that has further been summarized or derived from base warehouse ....

O'Neil, G. Graefe. Multi-Table Joins through Bitmapped Join Indices. SIGMOD Record,

....on the data warehouse team. Indexing generally falls within the role of the database administrator (DBA) but the design team is also necessary at this point to guide the DBA s decision. Two main indexing techniques used for data warehousing environments are bitmap indexes [BI98] and join indexes [OG95, Vald87]. A bitmap index is a B tree in which each leaf node is associated with a N bit string for N rows, where each bitmap stream is created for each value of the index. For example, a bitmap index for a gender attribute for 1 million customers will create two bitmap streams where the size of each is 1 ....

O'Neil, P. and Graefe, G., Multi-table Joins Through Bitmapped Join Indices, SIGMOD Record, Vol. 24, No.3, Sept. 1995, pp. 8-11.

....issue of the database community in the last years [3] To improve query performance in data warehouses (DWH) with a typically very high data volume, two major aspects must be considered. First, raw data access is accelerated by recently developed index and join techniques like bitwise joinindexing [10]. Second, pre aggregation of the raw data according to the need of the applications and their redundant storing in the warehouse offers a high potential of performance increase for OLAP queries [8] Another term for such a pre aggregate is materialized view (MV) which is used throughout this ....

O'Neil, P.; Graefe, G.: Multi-Table Joins Through Bitmapped Join Indices, SIGMOD Record 24(3), 1995

.... To support these kinds of queries, database vendors have significantly extended their query processors and researchers have just recently developed a large variety of new query processing techniques; e.g. the use of bitmap indices [CI98] special joins that exploit bitmap join indices [GO95] new join methods [HWM98] or multi query optimization for decision support [ZDNS98] to name just a few. In addition, a whole new industry, data warehouses, has appeared with products that materialize (i.e. pre compute) query results and cache the results of queries. Furthermore, the TPC H ....

G. Graefe and P. O'Neil. Multi-table joins through bitmapped join indices. ACM SIGMOD Record, 24(3):8--11, Oct 1995.

....tuples are inserted and deleted in the underlying relations. In situations where joins are taken often, the cost of doing this maintenance can be more than o#set by the savings achieved in performing the join. A number of commercial decision support systems are rumored to be using join indexes [O Neill and Graefe, 1995] In [Valduriez, 1987] Valduriez proposed and analyzed a join algorithm that uses the join index. The most important conclusion of that study was that, under many circumstances, having the join index allows one to compute the join significantly faster than the best ad hoc methods such as ....

O'Neill, P. and Graefe, G. (1995). Multi-table joins through bitmapped join indices. SIGMOD Record, 24(3).

....signatures for processing joins involving predicates on nested sets. Also, bitmap indexing is a well known concept; see, e.g. the early work on signature files [CS89] or the bitmap indices in Model 204 [O N87] Indexing attribute values via bitmaps [OQ97, CI98, WB98] and bitmap join indices [GO95] have recently received renewed attention in the context of query processing for data warehouses. To the best of our knowledge, however, so far nobody has used bitmaps for indirectly partitioning arguments of hash joins (or grouping operators) The most relevant related work are hash teams, ....

G. Graefe and P. O'Neil. Multi-table joins through bitmapped join indices. ACM SIGMOD Record, 24(3):8--11, Oct 1995.

....predicates can be efficiently evaluated directly on the bit vectors. Other Techniques Other indexing techniques for the warehouse environment include multidimensional B trees [8, 4] compression techniques (e.g. run length) for simple bitmap indexes, hierarchical indexes [6, 7] join indexes [15, 10] and multidimensional indexing for spatial data [12] Index techniques used in Sybase IQ, Red Brick Warehouse and Oracle please refer to [3] 5 Concluding Remarks and Future Work In this paper, we have introduced a new indexing technique, encoded bitmap indexing, for the DW environment. The ....

P. O'Neil, G. Graefe, Multi-Table Joins Through Bitmapped Join Indices, SIGMOD Record, Vol. 24, No. 3, September 1995.

....contain lists of record ids but bit vectors with one bit for each data record. The bit vector representation very efficiently supports the set operations such as union and intersection. For instance, OQ 97] discusses several types of bitmap index structures suitable for different query types. OG 95] introduces bitmap join indices which precompute binary joins in a data warehouse. Bitmap indices, however, are static because on the insertion of a data record all index entries have to be updated. Furthermore, one dimensional index structures build secondary indices which do not impact the ....

O'Neil P., Graefe G.: "Multi-Table Joins through Bitmapped Join Indices", SIGMOD Record 24(3), 1995, pp. 8-11.

No context found.

G. Graefe and P. O'Neil. Multi-table joins through bitmapped join indices. ACM SIGMOD Record, 24(3):8--11, October 1995.

No context found.

P. O'Neil and G. Graefe. Multi-table joins through bitmapped join indices. SIGMOD Record, 24(3):8--11, Sep 1995.

No context found.

P. E. O'Neil and G. Graefe. Multi-table joins through bitmapped join indices. SIGMOD Record, 24(3):8--11, 1995.

No context found.

P. E. O'Neil and G. Graefe. Multi-table joins through bitmapped join indices. SIGMOD Record, 24(3):8--11, 1995.

No context found.

P. E. O'Neil and G. Graefe. Multi-table joins through bitmapped join indices. SIGMOD Record, 24(3):8--11, 1995.

No context found.

P. O'Neil, G. Graefe, Multi-Table Joins Through Bitmapped Join Indices, SIGMOD Record, 24(3), September 1995.

No context found.

P. O'Neil, G. Graefe, Multi-Table Joins Through Bitmapped Join Indices, SIGMOD Record, 24(3), 1995.

No context found.

P. E. O'Neil and G. Graefe. Multi-table joins through bitmapped join indices. SIGMOD Record, 24(3):8--11, 1995.

No context found.

P.O. O'Neil and G. Graefe. Multi-Table Joins Through Bitmapped Join Indices. SIGMOD Record, 24(3), September 1995.

No context found.

G. Graefe and P. O'Neil. Multi-table joins through bitmapped join indices. ACM SIGMOD Record, 24(3):8--11, October 1995.

No context found.

P. O'Neil and G. Graefe. Multi-Table Joins Through Bitmapped Join Indices. ACM SIGMOD Record, pages 8--11, September 1995.

No context found.

Patrick O'Neil, Goetz Graefe. Multi-Table Joins Through Bitmapped Join Indices. SIGMOD Record, September 1995.

No context found.

P. O'Neil, G. Graefe, Multi-Table Joins Through Bitmapped Join Indices. SIGMOD Record, 8--11, September 1995.

First 50 documents  Next 50

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