<rt id="bn8ez"></rt>
<label id="bn8ez"></label>

  • <span id="bn8ez"></span>

    <label id="bn8ez"><meter id="bn8ez"></meter></label>

    Decode360's Blog

    業精于勤而荒于嬉 QQ:150355677 MSN:decode360@hotmail.com

      BlogJava :: 首頁 :: 新隨筆 :: 聯系 ::  :: 管理 ::
      302 隨筆 :: 26 文章 :: 82 評論 :: 0 Trackbacks
    1 . FISCAL YEAR TABLES
    ?
    ??? Let's write some CREATE TABLE statements that are as complete aspossible. This little exercise is important because SQL is a declarativelanguage and you need to learn how to specify things in the databaseinstead of in the code.
    ?
    ??? The table looks like this:
    ??? CREATE TABLE FiscalYearTable1
    ??? (fiscal_year INTEGER,
    ???? start_date?DATE,
    ???? end_date??? DATE);

    ?
    ??? It stores date ranges for determining what fiscal year any given datebelongs to. For example, the federal government runs its fiscal year fromOctober 1 until the end of September. The scalar subquery you woulduse to do this table lookup is:
    ??? (SELECT F1.fiscal_year
    ?????? FROM FiscalYearTable1 AS F1
    ????? WHERE outside_date BETWEEN F1.start_date AND F1.end_date)

    ?
    ??? Your assignment is to add all the constraints you can think of to thetable to guarantee that it contains only correct information.
    ?
    ??? While vendors all have different date and time functions, let's assumethat all we have is the SQL-92 temporal arithmetic and the function:
    ??? EXTRACT ([YEAR | MONTH | DAY] FROM <date expression>), whichreturns an integer that represents a field within a date.
    ?
    ?
    Answer #1
    ?
    ??? 1. First things first; make all the columns NOT NULL since thereis no good reason to allow them to be NULL.
    ?
    ??? 2. Most SQL programmers immediately think in terms of addinga PRIMARY KEY, so you might add the constraint ?PRIMARYKEY (fiscal_year, start_date, end_date) because the fiscal year isreally another name for the pair (start_date, end_date). This isnot enough, because it would allow this sort of error:
    ??? (1995, '1994-10-01', '1995-09-30')
    ??? (1996, '1995-10-01', '1996-08-30') <== error!
    ??? (1997, '1996-10-01', '1997-09-30')
    ??? (1998, '1997-10-01', '1997-09-30')

    ??? You could continue along the same lines and fix some problemsby adding the constraints UNIQUE (fiscal_year),UNIQUE (start_date), and UNIQUE (end_date), since we donot want duplicate dates in any of those columns.
    ?
    ??? 3. The constraint that almost everyone forgets to add because it isso obvious is:
    ??? CHECK (start_date < end_date) or CHECK (start_date <=end_date), as is appropriate.
    ?
    ??? 4. A better way would be to use the constraint PRIMARY KEY(fiscal_year) as before, but then since the start and end datesare the same within each year, you could use constraints onthose column declarations:
    ??? CREATE TABLE FiscalYearTable1
    ??? (fiscal_year INTEGER NOT NULL PRIMARY KEY,
    ???? start_date?DATE NOT NULL,
    ???? CONSTRAINT valid_start_date
    ???????CHECK ((EXTRACT (YEAR FROM start_date) = fiscal_year - 1)
    ?????????AND (EXTRACT (MONTH FROM start_date) = 10)
    ???????? AND (EXTRACT (DAY FROM start_date) = 01)),
    ???? end_date????DATE NOT NULL,
    ???? CONSTRAINT??valid_end_date
    ?????? CHECK ((EXTRACT (YEAR FROM end_date) = fiscal_year)
    ?????????AND (EXTRACT (MONTH FROM end_date) = 09)
    ???????? AND (EXTRACT (DAY FROM end_date) = 30)));

    ??? You could argue for making each predicate a separate constraintto give more detailed error messages. The predicates onthe year components of the start_date and end_date columnsalso guarantee uniqueness because they are derived from theunique fiscal year.
    ?
    ??? 5. Unfortunately, this method does not work for all companies.Many companies have an elaborate set of rules that involve takinginto account the weeks, weekends, and weekdays involved.They do this to arrive at exactly 360 days or 52 weeks in theiraccounting year. In fact, there is a fairly standard accountingpractice of using a “4 weeks, 4 weeks, 5 weeks” quarter withsome fudging at the end of the year; you can have a leftoverweek between 3 and 11 days. The answer is a FiscalMonthtable along the same lines as this FiscalYears example.
    ?
    ??? A constraint that will work surprisingly well for such cases is:
    ??? CHECK ((end_date - start_date) = INTERVAL 359 DAYS)
    ?
    ??? where you adjust the number of days to fit your rules (i.e., 52 weeks * 7?days = 364 days). If the rules allow some variation in the size of the fiscalyear, then replace the equality test with a BETWEEN predicate.
    ??? Now, true confession time. When I have to load such a table in adatabase, I get out my spreadsheet and build a table using the built-intemporal functions. Spreadsheets have much better temporal functionsthan databases, and there is a good chance that the accountingdepartment already has the fiscal calendar in a spreadsheet.
    ?
    ?




    -The End-

    posted on 2009-03-20 21:21 decode360-3 閱讀(192) 評論(0)  編輯  收藏 所屬分類: Toy
    主站蜘蛛池模板: 亚洲情A成黄在线观看动漫软件| 久久91亚洲人成电影网站| 亚洲国产成人va在线观看网址| 日韩精品免费在线视频| 久久久久久久尹人综合网亚洲| 伊人久久大香线蕉免费视频| 亚洲人成色777777在线观看| 在线观看特色大片免费网站| 亚洲AV无码一区二区乱子伦| 三年片在线观看免费观看大全动漫 | 免费91最新地址永久入口| 亚洲精品美女久久777777| 暖暖免费日本在线中文| 日产亚洲一区二区三区| 亚欧色视频在线观看免费| 亚洲人成图片网站| 国产美女无遮挡免费视频| 猫咪免费观看人成网站在线| 亚洲精品成人片在线观看| 久久久久久久国产免费看| 亚洲精品天天影视综合网| 免费在线视频你懂的| 亚洲无mate20pro麻豆| 国产大片91精品免费看3| jzzjzz免费观看大片免费| 亚洲av之男人的天堂网站| 国产免费不卡v片在线观看| 国产亚洲精品第一综合| 国产午夜亚洲精品午夜鲁丝片| 久久精品中文字幕免费| 亚洲精品伊人久久久久| 可以免费观看一级毛片黄a| 成人无码a级毛片免费| 亚洲一级片在线观看| 亚洲va中文字幕无码| 精品熟女少妇a∨免费久久| 国产精品亚洲一区二区在线观看| 亚洲精品乱码久久久久66| 国产一精品一AV一免费孕妇| 成年大片免费视频播放一级| 精品亚洲国产成AV人片传媒|