數(shù)據(jù)倉(cāng)庫(kù)的概念(更新版)
【正文】 istrict Time Location Type Area Occupant Name Purchase Price Rent ?. Sales Fact Table Type Size Area District Diatrict Region Region Region Building/ Estate Year Quarter Quarter Year Quarter Month Quarter Month Day Month Day Housing Types Time Location Type Area Occupant Name Purchase Price Rent ?. Sales Fact Table Day Month Quarter Year Time Territory District Region Building/Estate Geographic Location Type Size Area Housing Types Star schema Snowflake schema 數(shù)據(jù)倉(cāng)庫(kù)設(shè)計(jì)優(yōu)化的原則 ? 避免數(shù)據(jù)實(shí)時(shí)匯總 (建立匯總表 ) ? 減少表連接操作 (不要超過(guò) 35個(gè) ) ? 用 ID code作關(guān)鍵字 ? 減少 I/O競(jìng)爭(zhēng) ? 利用分區(qū)技術(shù)提高性能和可管理性 估算數(shù)據(jù)倉(cāng)庫(kù)容量的算法 X = size of one row in the fact table Therefore, Sales Fact Table = 98 Time Location Type Area Occupant Name Purchase Price Rent ?. Sales Fact Table Day Month Quarter Year Time Territory District Region Building/Estate Geographic Location Type Size Area Housing Types Building Property DW Design Dimension Estimated Time 96 month Housing Type Geographic Loc. 20 types 1000 locs. Y = sparsity (density factor) Estimated size of database = 98 * 96 * 20 * 1000 * = Mb 步驟 10: 從業(yè)務(wù)系統(tǒng)中抽取數(shù)據(jù)到數(shù)據(jù)倉(cāng)庫(kù) 數(shù)據(jù)抽取的要求： ? 可訪問(wèn)各種數(shù)據(jù)源 ? 可滿足時(shí)間要求 ? 可滿足數(shù)據(jù)轉(zhuǎn)換要求 ? 可檢測(cè)源系統(tǒng)中數(shù)據(jù)的變化 步驟 11: 開發(fā)前端應(yīng)用 步驟 12: 數(shù)據(jù)倉(cāng)庫(kù)的管理 ? 安全管理 ? 備份和恢復(fù) ? 高可用性 ? 數(shù)據(jù)時(shí)效 DEMO
教學(xué)課件相關(guān)推薦
鄂ICP備17016276號(hào)-1