logo, người, địa điểm, hoặc sự kiện là có chủ đích hoặc hàm ý. Microsoft tiêu đề báo chí có thể được mua để sử dụng quảng cáo giáo dục, kinh doanh, hoặc bán hàng. Phiên bản trực tuyến cũng có sẵn cho hầu hết | 58 Chapter 3 Getting to Trusted Data Before cleansing data your primary goal is to determine where the data is to extract clean and conform. Because the ETL process can be the most resource heavy it s also vital that you understand exactly what the users require before you begin. Best practices for understanding user requirements often include a prototyping and iterative back-and-forth discussion among analysts report users and developers. It is particularly important to validate that you re collecting the right amount of information at the right time in the right format and with the most helpful visualizations. Far too often a company starting a BI initiative wants to integrate the enterprise s data right now so they re reluctant to take the time to design an extensible and universal solution but not doing so increases the likelihood of unusable solutions that lead to future rework. It is well worth your time to review the methodologies and literature around BI project management solutions in other books and articles. You deliver the cleaned source data into a dimensional data store that implements querying and analysis capabilities for the purpose of decision-making. There are excellent books and articles on best practices for ETL data. Ralph Kimball is a pioneer for data warehousing and has published much of the original content for ETL and other data warehousing best practices. You can find his books and articles at http . Other sources include The Data Warehouse Institute TDWI at http which is an educational institution for BI and data warehousing. In addition the Information Management site at http www. provides news commentary and feature content serving the information technology and business community. Step 3 Create and Populate the Data Warehouse As the following illustration shows Step 3 occurs in two phases. First you must design and create a data warehouse in SQL Server. A number of books have .
