Question

我有一个名为user_transaction的数据库表，其结构如下：

transaction_id    mediumint(6)   UNSIGNED (PK)
transaction_no    varchar(55)
transaction_cc_avenue_no    varchar(55)
transaction_card_category    varchar(100)
transaction_user_id      varchar(32)
transaction_user_name    varchar(255)
transaction_user_email_id    varchar(255)
transaction_deal_code    varchar(10)
transaction_dc_id    smallint(4)
transaction_amount    float(10,2)
transaction_discount    float(10,2)
transaction_total_amount    float(10,2)
transaction_data_assign    enum('0', '1')
transaction_status    enum('success', 'inprocess', 'fail', 'cancelled')     
transaction_date    bigint(12)
transaction_update_date    bigint(12)
transaction_update_user_id    varchar(32)

我正在使用UNIX时间戳值将日期存储在transaction_date列中。现在我以“dd/mm/yyyy”格式从表单中提取两个日期，在UNIX时间戳中将其转换并在以下查询中使用它。现在我要显示日期号码。不同transaction_status的交易（即总交易发生在每个日期，状态'成功'，'进行中'，'失败'和'取消'的交易总数）对于我给出的范围之间的所有日期。这些记录应按交易日期分组。我尝试了很多以获得这个结果，但没有取得任何成功。供你参考我在下面提出我的问题：

SELECT COUNT(*) `total count`, SUM(transaction_status = 'success') `success`, SUM(transaction_status = 'inprocess') `inprocess`, SUM(transaction_status = 'fail') `fail`, SUM(transaction_status = 'cancelled') `cancelled` FROM user_transaction WHERE transaction_date >= '1325376000' AND transaction_date <=  '1338422400' GROUP BY FROM_UNIXTIME(transaction_date)

在这方面有谁能帮助我？提前谢谢。

Answer 1

from_unixtime()功能会创建一个带有时间的日期。你只想要日期：

SELECT date(FROM_UNIXTIME(transaction_date)),
       COUNT(*) `total count`, SUM(transaction_status = 'success') `success`, 
       SUM(transaction_status = 'inprocess') `inprocess`,
       SUM(transaction_status = 'fail') `fail`,
       SUM(transaction_status = 'cancelled') `cancelled`
FROM user_transaction
WHERE transaction_date >= 1325376000 AND transaction_date <= 1338422400
GROUP BY date(FROM_UNIXTIME(transaction_date))

我还从数字常量中删除了'。这些是数字，因此使它们看起来像字符串是错误的（不影响性能，只是查询的可理解性）。

编辑：

如果您想按月分组：

SELECT year(date(FROM_UNIXTIME(transaction_date))),
       month(date(FROM_UNIXTIME(transaction_date))),
       COUNT(*) `total count`, SUM(transaction_status = 'success') `success`, 
       SUM(transaction_status = 'inprocess') `inprocess`,
       SUM(transaction_status = 'fail') `fail`,
       SUM(transaction_status = 'cancelled') `cancelled`
FROM user_transaction
WHERE transaction_date >= 1325376000 AND transaction_date <= 1338422400
GROUP BY year(date(FROM_UNIXTIME(transaction_date))),
         month(date(FROM_UNIXTIME(transaction_date)))

您可能需要更改所包含日期范围的where子句。

如何显示以下方案的分组结果？

1 个答案: