Academic journal article American Academic & Scholarly Research Journal

The Bitmap Index Advantages on the Data Warehouses

Academic journal article American Academic & Scholarly Research Journal

The Bitmap Index Advantages on the Data Warehouses

Article excerpt

Abstract. The data warehouse designer should consider the effectiveness of the data query while the selection of relevant indexes and their combination with materialized views, this, because the data query is the only way to get this information from a data warehouse. The index selection is considered an NP-complete problem, because of the number of indexes is exponential in the total attributes in the data warehouse tables. So, the choose of the suitable index type is considered as the main step to start the data warehouse design. In this paper we are making a comparison study between the B-tree index as traditional index type, and the Bitmap index. this comparison are based on three factors : index size, clustering factor and compression, and we illustrate this with a real experiment.

Keywords: Business Intelligence, Data warehouse.

1 INTRODUCTION

Before the data warehouse implementation, some administration tasks that are taken by a data warehouse administrator needs to be decided, like logical and physical design, management of storage space and performance tuning.

The most important task is the physical design including data organization and access improvement. The fast access to this data needs general index to find the information wanted without reviewing all table data.

As per this need, the index selection task is considered difficult because their number is exponential in the total number of attributes in the database. So the index plays an important role in the data warehouses performance. For this reason, we focus on this data warehouse aspect, which is considered interesting for the designer while editing and query optimization selection.

The target is to minimize the query execution time, and as in a data warehouse the query uses indexes to access to the data, we will work on the problem of choosing the type of index when designing our data warehouse.

It's also important, that the index compression and size must be checked to optimize the storage usage.

There are several types of indexes supported by databases such as Bitmap, B-tree, Bitmap join, range-based bitmap index etc.. In this sense we have chosen two types of index relevant to this study, the index type: B-tree index and type Bitmap.

2 BITMAP INDEX

2.1 Definition

A bitmap index is a data structure defined in a DBMS used to optimize access to data in databases. It is a type of indexing is particularly interesting and effective in the context of selection queries. The index bitmap attribute is encoded in bits, where its low cost in terms of space occupied. All possible attribute values are considered, the value is present or not in the table. Each of these values is an array of bits, called bitmap, which contains as many bits as ntuples present in the table. Thus, this type of index is very effective when the attributes have a low number of distinct values. Each bit represents the value of an attribute for a given tuple. For each bit, there is an encoding presence / absence (1/0), which indicates that a tuple or not the present value characterized in bitmap.

To illustrate how a bitmap index works, we take an example EE-PP-O'Neil and O'Neil, Table 1 illustrates a basic bitmap index into a table containing 9 records, where the index is created in the C column with integers ranging from 0 to 3, we say that the cardinality of the column C is 4, by what there are 4 distinct values [0, 1, 2, 3], where the index bitmap C Contains 4 bitmaps shown as B0, B1, B2 and B3 corresponding value represents. In this example, the first line where RowID = 0, column C is worth 2, consequently, B2 column bit value "1", while the other bitmaps are set to "0". Same for the next line, where C = 1 corresponds to the bitmap B1 is set to 1 and the rest to "0". This process is repeated for the remaining lines.

2.1 Properties

Bitmap indexes have a very interesting property of responding to certain types of requests without returning the data themselves, thus optimizing the response time, disk storage. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.