sql - 列中的按月总计

Question

我有一张表（消费）如下：

DocDate   ItemCode    Quantity
01.01.09  A           5
02.01.09  A           6

等等..全年

我需要如下结果：

ItemCode Jan09Total Feb09Total Mar09Total

问题是应该对月份的数量进行汇总，这些数量应该在列中并根据给定的日期标准，例如从 1 月到 3 月 9 日或 2 月到 7 月 9 日等。

如何使用递归查询来实现这一点。

谢谢拉胡尔

Answer 1

Assuming you use a DB which supports PIVOT, you should use it. Here's example code for Oracle:

CREATE TABLE Consumption (
  DocDate  DATE,
  ItemCode VARCHAR2(10),
  Quantity NUMBER
);

INSERT INTO Consumption VALUES(to_date('2009-01-01', 'YYYY-MM-DD'), 'A', 5);
INSERT INTO Consumption VALUES(to_date('2009-02-01', 'YYYY-MM-DD'), 'A', 6);
INSERT INTO Consumption VALUES(to_date('2009-01-01', 'YYYY-MM-DD'), 'B', 5);

SELECT * FROM Consumption
PIVOT(
  sum(Quantity)
  FOR DocDate IN (
    to_date('2009-01-01', 'YYYY-MM-DD') AS Jan09Total,
    to_date('2009-02-01', 'YYYY-MM-DD') AS Feb09Total
  )
);

As noted by Robin, if your columns vary, then you will have to dynamically generate such a query with the correct columns.

Answer 2

好吧，它不是递归的，但这会让你得到你想要的：

SELECT
    DateName(month, DocDate) + ' ' + CAST(YEAR(DocDate) AS VARCHAR(10)) AS MonthYear,
    ItemCode,
    SUM(Quantity)
FROM
    Consumption
GROUP BY
    DateName(month, DocDate) + ' ' + CAST(YEAR(DocDate) AS VARCHAR(10)) AS MonthYear,
    ItemCode
ORDER BY
    1,2

我认为递归只会阻碍这个查询的性能。你确定要这样做吗？或者这只是家庭作业？

Answer 3

Here is a link to a wonderfully elegant way of implementing pivot logic on KNOWN x-axis values. Dynamic pivots are more difficult.

I didn't read this article fully, so if it isn't helpful then look up "The Rozenshtein Method"

http://www.stephenforte.net/default.aspx?date=2003-08-07

I've used this in many queries with great results.

Answer 4

您可以根据需要使用 PIVOT 将数据放入列中，但是，需要事先定义列列表。因此，您需要为您需要的列（基于您的日期范围）构建一些动态 SQL，然后运行它来生成您的 PIVOT。