如何填充新列？

Question

我向已经有很多行的 DailyTotals table 添加了四个新列，我需要通过查询所有每日典当交易，用每行的每日总计填充新列。

DailyTotals table 看起来像这样：

    tblDailyTotals
    (
    dt_date DATE,       {unique index}
    ...
    dt_P_count INT,
    dt_P_amt DECIMAL(8,2),
    dt_X_count INT,
    dt_X_amt DECIMAL(8,2)
    );

典当交易 table 看起来像这样：

    tblPawnTransactions
    (
    ...
    pt_date DATE,       {dups index}
    pt_type CHAR(1), {"P","X"}
    pt_amt DECIMAL(7,2)
    );

您将如何编写批量更新查询代码以填充 DailyTotals 中每一天的四个新列 table？

Answer 1

这是一些可用的 Informix SQL。它假定您在环境中设置了 DBDATE='Y4MD-' 或等效设置——您需要调整日期格式以适合您的偏好。

设置

CREATE TABLE    tblDailyTotals
    (
    dt_date DATE PRIMARY KEY NOT NULL,       {unique index}
    dt_other CHAR(20) NOT NULL,
    dt_P_count INT,
    dt_P_amt DECIMAL(8,2),
    dt_X_count INT,
    dt_X_amt DECIMAL(8,2)
    );


CREATE TABLE   tblPawnTransactions
    (
    pt_customer INT,
    pt_date DATE,       {dups index}
    pt_type CHAR(1), {"P","X"}
    pt_amt DECIMAL(7,2)
    );

INSERT INTO tblDailyTotals(dt_date, dt_other) VALUES('2017-01-01', 'Other info 2017-01-01');
INSERT INTO tblDailyTotals(dt_date, dt_other) VALUES('2017-01-02', 'Other info 2017-01-02');
INSERT INTO tblDailyTotals(dt_date, dt_other) VALUES('2017-01-03', 'Other info 2017-01-03');
INSERT INTO tblDailyTotals(dt_date, dt_other) VALUES('2017-01-04', 'Other info 2017-01-04');
INSERT INTO tblDailyTotals(dt_date, dt_other) VALUES('2017-01-05', 'Other info 2017-01-05');
INSERT INTO tblDailyTotals(dt_date, dt_other) VALUES('2017-01-06', 'Other info 2017-01-06');
UPDATE tblDailyTotals
   SET dt_P_count = 3, dt_P_amt = 345.67, dt_X_count = 9, dt_X_amt = 291.23
 WHERE dt_date = '2017-01-06';

INSERT INTO tblPawnTransactions VALUES(1, '2017-01-01', 'P', 12.34);
INSERT INTO tblPawnTransactions VALUES(2, '2017-01-01', 'P', 22.25);
INSERT INTO tblPawnTransactions VALUES(3, '2017-01-01', 'X', 32.16);
INSERT INTO tblPawnTransactions VALUES(1, '2017-01-02', 'P', 42.07);
INSERT INTO tblPawnTransactions VALUES(4, '2017-01-02', 'P', 52.98);
INSERT INTO tblPawnTransactions VALUES(2, '2017-01-02', 'X', 62.89);
INSERT INTO tblPawnTransactions VALUES(1, '2017-01-03', 'P', 72.70);
INSERT INTO tblPawnTransactions VALUES(1, '2017-01-03', 'X', 82.51);
INSERT INTO tblPawnTransactions VALUES(1, '2017-01-03', 'P', 92.42);
INSERT INTO tblPawnTransactions VALUES(2, '2017-01-05', 'X', 13.33);
INSERT INTO tblPawnTransactions VALUES(1, '2017-01-05', 'P', 14.14);
INSERT INTO tblPawnTransactions VALUES(1, '2017-01-05', 'X', 15.05);

更新

SELECT * FROM tblDailyTotals ORDER BY dt_date;

UPDATE tblDailyTotals
   SET (dt_P_count, dt_P_amt) = ((SELECT COUNT(*), SUM(pt_amt)
                                    FROM tblPawnTransactions AS pt
                                   WHERE pt.pt_date = tblDailyTotals.dt_date
                                     AND pt.pt_type = 'P'
                                ))
 WHERE EXISTS(SELECT * FROM tblPawnTransactions AS pt
               WHERE pt.pt_date = tblDailyTotals.dt_date AND pt_type = 'P')
;

UPDATE tblDailyTotals
   SET (dt_X_count, dt_X_amt) = ((SELECT COUNT(*), SUM(pt_amt)
                                    FROM tblPawnTransactions AS pt
                                   WHERE pt.pt_date = tblDailyTotals.dt_date
                                     AND pt.pt_type = 'X'
                                ))
 WHERE EXISTS(SELECT * FROM tblPawnTransactions AS pt
               WHERE pt.pt_date = tblDailyTotals.dt_date AND pt_type = 'X')
;

双括号是必须的；外对表示 SET 子句的 RHS 上有一个值列表，以匹配 SET 子句的 LHS 上括号中的列列表，而内括号表示值是 [= 的结果41=]声明。

示例数据的输出

更新前：

2017-01-01  Other info 2017-01-0
2017-01-02  Other info 2017-01-0
2017-01-03  Other info 2017-01-0
2017-01-04  Other info 2017-01-0
2017-01-05  Other info 2017-01-0
2017-01-06  Other info 2017-01-0    3   345.67      9  291.23

更新后：

2017-01-01  Other info 2017-01-0    2   34.59       1   32.16
2017-01-02  Other info 2017-01-0    2   95.05       1   62.89
2017-01-03  Other info 2017-01-0    2   165.12      1   82.51
2017-01-04  Other info 2017-01-0
2017-01-05  Other info 2017-01-0    1   14.14       2   28.38
2017-01-06  Other info 2017-01-0    3   345.67      9  291.23

为什么使用 EXISTS 子句？

如果省略 WHERE EXISTS(…) 子句会怎样？ tblPawnTransactions table 中没有匹配条目的行也会更新，如下所示：

SELECT * FROM tblDailyTotals ORDER BY dt_date;

UPDATE tblDailyTotals
   SET (dt_P_count, dt_P_amt) = ((SELECT COUNT(*), SUM(pt_amt)
                                    FROM tblPawnTransactions AS pt
                                   WHERE pt.pt_date = tblDailyTotals.dt_date
                                     AND pt.pt_type = 'P'
                                ))
 ;

UPDATE tblDailyTotals
   SET (dt_X_count, dt_X_amt) = ((SELECT COUNT(*), SUM(pt_amt)
                                    FROM tblPawnTransactions AS pt
                                   WHERE pt.pt_date = tblDailyTotals.dt_date
                                     AND pt.pt_type = 'X'
                                ))
 ;

SELECT * FROM tblDailyTotals ORDER BY dt_date;

这会生成：

2017-01-01  Other info 2017-01-0    2   34.59       1   32.16
2017-01-02  Other info 2017-01-0    2   95.05       1   62.89
2017-01-03  Other info 2017-01-0    2   165.12      1   82.51
2017-01-04  Other info 2017-01-0    0               0
2017-01-05  Other info 2017-01-0    1   14.14       2   28.38
2017-01-06  Other info 2017-01-0    0               0

请注意，2017-01-04 和 2017-01-06 的行都已更新。现在，如果您希望这些零和空值就位，那很好。如果不是，请使用 EXISTS 子句。

如何填充新列？

How can I populate new columns?

sql

informix

bulkupdate

设置

更新

示例数据的输出

为什么使用 EXISTS 子句？