Sql 当 A = Z 则设置 B = X
Sql When A = Z then set B = X
我似乎无法运行此查询，它一直告诉我“消息 102，级别 15，状态 1，第 5 行
'=' 附近的语法不正确。”不确定我在这里做错了什么。
基本上我需要 [forwarder_preferred_status] 为 P2 或 P3，具体取决于 [forwarder_short]
的值
欢迎提出任何建议
Select
[forwarder_display],
Case [forwarder_preferred_status]
-- Update P2 Customers
When [forwarder_short] = 'ABL' THEN [forwarder_preferred_status] = 'P2'
When [forwarder_short] = 'ALK' THEN [forwarder_preferred_status] = 'P2'
When [forwarder_short] = 'EIF' THEN [forwarder_preferred_status] = 'P2'
When [forwarder_short] = 'NNR' THEN [forwarder_preferred_status] = 'P2'
When [forwarder_short] = 'ALI' THEN [forwarder_preferred_status] = 'P2'
When [forwarder_short] = 'LAF' THEN [forwarder_preferred_status] = 'P2'
When [forwarder_short] = 'AIT' THEN [forwarder_preferred_status] = 'P2'
-- Update P3 Customers
When [forwarder_short] = 'SCHBAX' THEN [forwarder_preferred_status] = 'P3'
When [forwarder_short] = 'UPS' THEN [forwarder_preferred_status] = 'P3'
When [forwarder_short] = 'KUE' THEN [forwarder_preferred_status] = 'P3'
When [forwarder_short] = 'AGI' THEN [forwarder_preferred_status] = 'P3'
When [forwarder_short] = 'PAN' THEN [forwarder_preferred_status] = 'P3'
When [forwarder_short] = 'DGF' THEN [forwarder_preferred_status] = 'P3'
When [forwarder_short] = 'NEC' THEN [forwarder_preferred_status] = 'P3'
When [forwarder_short] = 'HWE' THEN [forwarder_preferred_status] = 'P3'
When [forwarder_short] = 'NIS' THEN [forwarder_preferred_status] = 'P3'
When [forwarder_short] = 'BDP' THEN [forwarder_preferred_status] = 'P3'
When [forwarder_short] = 'KWE' THEN [forwarder_preferred_status] = 'P3'
When [forwarder_short] = 'CEVA' THEN [forwarder_preferred_status] = 'P3'
When [forwarder_short] = 'VIZ' THEN [forwarder_preferred_status] = 'P3'
When [forwarder_short] = 'FTN' THEN [forwarder_preferred_status] = 'P3'
When [forwarder_short] = 'UTI' THEN [forwarder_preferred_status] = 'P3'
-- Fix wrong prefered status
When [forwarder_short] = 'OIA' THEN [forwarder_preferred_status] = null
When [forwarder_short] = 'PHO' THEN [forwarder_preferred_status] = null
When [forwarder_short] = 'REI' THEN [forwarder_preferred_status] = null
Else [forwarder_preferred_status]
End as [forwarder_preferred_status]
Sum ([calc_gross_rev]) As [Gr Revenue]
Sum ([charge_weight]) As [Ch Weight]
From
[dbo].[report_bo_awb_revenue_all]
Where
[yr] = 2014
AND [agent_iata_code] not in ('0508634','0514616')
您不需要在 CASE 语句的 THEN 部分指定字段名称，您可以在比较完成时删除第一个 WHEN 之前的语法在每个 WHEN 子句中：
CASE WHEN [forwarder_short] = 'ABL' THEN 'P2'...
End as [forwarder_preferred_status]
另一种方法是使用 IN 子句来保存一些字符：
CASE WHEN [forwarder_short] IN('ABL','ALK','EIF'...) THEN 'P2'...
End as [forwarder_preferred_status]
您使用 as（或在表达式前使用 =）在 case 之后为 添加别名 。像这样：
(Case When [forwarder_short] = 'ABL' THEN 'P2'
When [forwarder_short] = 'ALK' THEN 'P2'
. . .
End) as forwarder_preferred_status
或（Aaron Bertrand 喜欢但我不喜欢）：
forwarder_preferred_status = (Case When [forwarder_short] = 'ABL' THEN 'P2'
When [forwarder_short] = 'ALK' THEN 'P2'
. . .
End);
另外，您在case后的表达式与when中布尔表达式的使用不一致。这只是坚持使用布尔表达式的情况的形式。
您可以使用 in:
大大简化您的逻辑
(case when forwarder_short in ('ABL', 'ALK', 'EIF', 'NNR', 'ALI', 'LAF', 'AIT')
then 'P2'
. . .
end) as forwarder_preferred_status
你问题中的代码结构如下：
--Bad
CASE A
WHEN X THEN A = B
WHEN Y THEN A = C
...
END
相反，您需要像这样构造它：
--Good
A = CASE
WHEN X THEN B
WHEN Y THEN C
...
END
或者像这样：
--Good
CASE
WHEN X THEN B
WHEN Y THEN C
...
END As A
case 语句仅产生 值，而不是表达式。它不会在任意代码之间分支。此外，假设您有很多代码产生相同的值，您可以大大 简化这样的事情：
--Better
CASE
WHEN X IN (1,2,3) THEN B
WHEN X IN (4,5,6) THEN C
...
ELSE NULL
END
看起来像这样：
Select
[forwarder_display],
CASE
WHEN [forwarder_short] IN ('ABL', 'ALK', 'EIF', 'NNR', 'ALI', 'LAF', 'AIT')
THEN 'P2'
WHEN [forwarder_short] IN ('SCHBAX', 'UPS', 'KUE', 'AGI', 'PAN', 'DGF', 'NEC',
'HWE', 'NIS', 'BDP', 'KWE', 'CEVA', 'VIZ', 'FTN', 'UTI')
THEN 'P3'
WHEN [forwarder_short] IN ('OIA','PHO', 'REI')
THEN NULL
ELSE [forwarder_preferred_status]
END as [forwarder_preferred_status],
Sum ([calc_gross_rev]) As [Gr Revenue],
Sum ([charge_weight]) As [Ch Weight]
From
[dbo].[report_bo_awb_revenue_all]
Where
[yr] = 2014
AND [agent_iata_code] not in ('0508634','0514616')
不幸的是，这仍然不起作用，因为您缺少使 SUM() 聚合函数起作用所需的 GROUP BY 子句。由于表达式复杂，我发现通过嵌套查询更容易解决问题，如下所示：
SELECT forwarder_display, [forwarder_preferred_status],
SUM([calc_gross_rev]) As [Gr Revenue],
SUM([charge_weight]) As [Ch Weight]
FROM
(
Select
[forwarder_display],
CASE
WHEN [forwarder_short] IN ('ABL', 'ALK', 'EIF', 'NNR', 'ALI', 'LAF', 'AIT')
THEN 'P2'
WHEN [forwarder_short] IN ('SCHBAX', 'UPS', 'KUE', 'AGI', 'PAN', 'DGF', 'NEC',
'HWE', 'NIS', 'BDP', 'KWE', 'CEVA', 'VIZ', 'FTN', 'UTI')
THEN 'P3'
WHEN [forwarder_short] IN ('OIA','PHO', 'REI')
THEN NULL
ELSE [forwarder_preferred_status]
END as [forwarder_preferred_status],
[calc_gross_rev],
[charge_weight]
From
[dbo].[report_bo_awb_revenue_all]
Where
[yr] = 2014
AND [agent_iata_code] not in ('0508634','0514616')
) t
GROUP BY [forwarder_display], [forwarder_preferred_status]
这应该 运行 并给你想要的结果，除非出现一些错误（很有可能，因为我在答案 window 中输入了所有这些未经测试的内容）。但是，您可能会发现仅通过 copy/pasting 对 GROUP BY 子句的复杂 CASE 表达式就可以获得更好的执行计划，而无需嵌套查询。
最后，像这样的代码——所有这些 3 位和 4 位代码都被硬编码了——要求将这些代码放入 table，以便通过 JOIN + 完成COALESCE，而不是 CASE。这将 MUCH 提高性能和维护。那将允许你写这样的东西：
--Best
SELECT [forwarder_display],
COALESCE(M.status, A.forwarder_preferred_status) As forwarder_preferred_status
Sum ([calc_gross_rev]) As [Gr Revenue],
Sum ([charge_weight]) As [Ch Weight]
FROM [dbo].[report_bo_awb_revenue_all] A
LEFT JOIN forwarder_map M ON m.forwarder_short = A.forwarder_short
WHERE
A.[yr] = 2014
AND [agent_iata_code] not in ('0508634','0514616')
GROUP BY [forwarder_display], COALESCE(M.status, A.forwarder_preferred_status)
最后一点：要使上面的示例起作用，您需要使用空字符串而不是虚构 forwarder_map table 中的 NULL 值。然后，您可以使用 NULLIF() 取回 NULL，如果这确实是您需要的。
你的陈述很难read/edit...你可以使用shorter/simpler方式..像这样：
SELECT [forwarder_display],
CASE
WHEN [forwarder_short] IN ('ABL','ALK','NNR','EIF','ALI','LAF','AIT') THEN 'P2'
WHEN [forwarder_short] IN ('SCHBAX', 'UPS', 'KUE','AGI','PAN'......) THEN 'P3'
WHEN [forwarder_short] IN ('OIA','PHO','REI') THEN null
ELSE [forwarder_preferred_status]
END AS forwarder_preferred_status
...
我敢打赌，这不是第一次有人需要将 ABL 翻译成 P2 来制作报告，也不会是最后一次。正如@Aaron 指出的那样，停止这些可怕的案例陈述并将其表格化。
为了这个答案，我使用了一个 table 变量。这些通常是不可取的，因为数据库引擎不会记录任何关于其中的值及其分布的统计信息，这是制定良好查询计划的关键。
DECLARE
@TRANSLATE table
(
forwarder_short varchar(10) NOT NULL
, forwarder_preferred_status varchar(10) NULL
);
INSERT INTO
@TRANSLATE
(
forwarder_short
, forwarder_preferred_status
)
VALUES
('ABL', 'P2')
, ('ALK', 'P2')
, ('EIF', 'P2')
, ('NNR', 'P2')
, ('ALI', 'P2')
, ('LAF', 'P2')
, ('AIT', 'P2')
, ('SCHBAX', 'P3')
, ('UPS', 'P3')
, ('KUE', 'P3')
, ('AGI', 'P3')
, ('PAN', 'P3')
, ('DGF', 'P3')
, ('NEC', 'P3')
, ('HWE', 'P3')
, ('NIS', 'P3')
, ('BDP', 'P3')
, ('KWE', 'P3')
, ('CEVA', 'P3')
, ('VIZ', 'P3')
, ('FTN', 'P3')
, ('UTI', 'P3')
, ('OIA', NULL)
, ('PHO', NULL)
, ('REI', NULL);
现在我有一个 table、TRANSLATE，它提供了我们已知的 forwarder_short 值到汇总 forwarder_preferred_status 之间的映射。
为了涵盖 forwarder_short 值不在此 table 中的情况，我们需要使用我们的翻译 table 作为 OUTER JOIN。
WITH report_bo_awb_revenue_all AS
(
-- simulating your source data
SELECT
T.forwarder_short
, T.forwarder_preferred_status
, 100 AS calc_gross_rev
, 12 AS charge_weigh
, 2014 AS yr
, '' AS agent_iata_code
, 'fwd' AS forwarder_display
FROM
@TRANSLATE AS T
UNION ALL
SELECT
'UNK'
, 'P90x'
, 9000 AS calc_gross_rev
, 12 AS charge_weigh
, 2014 AS yr
, '' AS agent_iata_code
, 'fwd' AS forwarder_display
)
SELECT
R.forwarder_display
, COALESCE(T.forwarder_preferred_status, R.forwarder_preferred_status) AS forwarder_preferred_status
, SUM(R.calc_gross_rev) AS calc_gross_rev
, SUM(R.charge_weigh) AS charge_weigh
FROM
report_bo_awb_revenue_all AS R
LEFT OUTER JOIN
@TRANSLATE AS T
ON T.forwarder_short = R.forwarder_short
WHERE
R.yr = 2014
AND R.agent_iata_code not in ('0508634','0514616')
GROUP BY
R.forwarder_display
, T.forwarder_preferred_status
, R.forwarder_preferred_status;
现在，所有的业务逻辑都被提炼成了这个
, COALESCE(T.forwarder_preferred_status, R.forwarder_preferred_status) AS forwarder_preferred_status
"Use the value in our translate table if it exists, otherwise, use what's stored on the report_bo_awb_revenue_all table."
SQLFiddle version 这使用临时 table 进行翻译，因为 fiddle 不处理 table 变量，但核心概念仍然存在。
我似乎无法运行此查询，它一直告诉我“消息 102，级别 15，状态 1，第 5 行 '=' 附近的语法不正确。”不确定我在这里做错了什么。
基本上我需要 [forwarder_preferred_status] 为 P2 或 P3，具体取决于 [forwarder_short]
的值欢迎提出任何建议
Select
[forwarder_display],
Case [forwarder_preferred_status]
-- Update P2 Customers
When [forwarder_short] = 'ABL' THEN [forwarder_preferred_status] = 'P2'
When [forwarder_short] = 'ALK' THEN [forwarder_preferred_status] = 'P2'
When [forwarder_short] = 'EIF' THEN [forwarder_preferred_status] = 'P2'
When [forwarder_short] = 'NNR' THEN [forwarder_preferred_status] = 'P2'
When [forwarder_short] = 'ALI' THEN [forwarder_preferred_status] = 'P2'
When [forwarder_short] = 'LAF' THEN [forwarder_preferred_status] = 'P2'
When [forwarder_short] = 'AIT' THEN [forwarder_preferred_status] = 'P2'
-- Update P3 Customers
When [forwarder_short] = 'SCHBAX' THEN [forwarder_preferred_status] = 'P3'
When [forwarder_short] = 'UPS' THEN [forwarder_preferred_status] = 'P3'
When [forwarder_short] = 'KUE' THEN [forwarder_preferred_status] = 'P3'
When [forwarder_short] = 'AGI' THEN [forwarder_preferred_status] = 'P3'
When [forwarder_short] = 'PAN' THEN [forwarder_preferred_status] = 'P3'
When [forwarder_short] = 'DGF' THEN [forwarder_preferred_status] = 'P3'
When [forwarder_short] = 'NEC' THEN [forwarder_preferred_status] = 'P3'
When [forwarder_short] = 'HWE' THEN [forwarder_preferred_status] = 'P3'
When [forwarder_short] = 'NIS' THEN [forwarder_preferred_status] = 'P3'
When [forwarder_short] = 'BDP' THEN [forwarder_preferred_status] = 'P3'
When [forwarder_short] = 'KWE' THEN [forwarder_preferred_status] = 'P3'
When [forwarder_short] = 'CEVA' THEN [forwarder_preferred_status] = 'P3'
When [forwarder_short] = 'VIZ' THEN [forwarder_preferred_status] = 'P3'
When [forwarder_short] = 'FTN' THEN [forwarder_preferred_status] = 'P3'
When [forwarder_short] = 'UTI' THEN [forwarder_preferred_status] = 'P3'
-- Fix wrong prefered status
When [forwarder_short] = 'OIA' THEN [forwarder_preferred_status] = null
When [forwarder_short] = 'PHO' THEN [forwarder_preferred_status] = null
When [forwarder_short] = 'REI' THEN [forwarder_preferred_status] = null
Else [forwarder_preferred_status]
End as [forwarder_preferred_status]
Sum ([calc_gross_rev]) As [Gr Revenue]
Sum ([charge_weight]) As [Ch Weight]
From
[dbo].[report_bo_awb_revenue_all]
Where
[yr] = 2014
AND [agent_iata_code] not in ('0508634','0514616')
您不需要在 CASE 语句的 THEN 部分指定字段名称，您可以在比较完成时删除第一个 WHEN 之前的语法在每个 WHEN 子句中：
CASE WHEN [forwarder_short] = 'ABL' THEN 'P2'...
End as [forwarder_preferred_status]
另一种方法是使用 IN 子句来保存一些字符：
CASE WHEN [forwarder_short] IN('ABL','ALK','EIF'...) THEN 'P2'...
End as [forwarder_preferred_status]
您使用 as（或在表达式前使用 =）在 case 之后为 添加别名 。像这样：
(Case When [forwarder_short] = 'ABL' THEN 'P2'
When [forwarder_short] = 'ALK' THEN 'P2'
. . .
End) as forwarder_preferred_status
或（Aaron Bertrand 喜欢但我不喜欢）：
forwarder_preferred_status = (Case When [forwarder_short] = 'ABL' THEN 'P2'
When [forwarder_short] = 'ALK' THEN 'P2'
. . .
End);
另外，您在case后的表达式与when中布尔表达式的使用不一致。这只是坚持使用布尔表达式的情况的形式。
您可以使用 in:
(case when forwarder_short in ('ABL', 'ALK', 'EIF', 'NNR', 'ALI', 'LAF', 'AIT')
then 'P2'
. . .
end) as forwarder_preferred_status
你问题中的代码结构如下：
--Bad
CASE A
WHEN X THEN A = B
WHEN Y THEN A = C
...
END
相反，您需要像这样构造它：
--Good
A = CASE
WHEN X THEN B
WHEN Y THEN C
...
END
或者像这样：
--Good
CASE
WHEN X THEN B
WHEN Y THEN C
...
END As A
case 语句仅产生 值，而不是表达式。它不会在任意代码之间分支。此外，假设您有很多代码产生相同的值，您可以大大 简化这样的事情：
--Better
CASE
WHEN X IN (1,2,3) THEN B
WHEN X IN (4,5,6) THEN C
...
ELSE NULL
END
看起来像这样：
Select
[forwarder_display],
CASE
WHEN [forwarder_short] IN ('ABL', 'ALK', 'EIF', 'NNR', 'ALI', 'LAF', 'AIT')
THEN 'P2'
WHEN [forwarder_short] IN ('SCHBAX', 'UPS', 'KUE', 'AGI', 'PAN', 'DGF', 'NEC',
'HWE', 'NIS', 'BDP', 'KWE', 'CEVA', 'VIZ', 'FTN', 'UTI')
THEN 'P3'
WHEN [forwarder_short] IN ('OIA','PHO', 'REI')
THEN NULL
ELSE [forwarder_preferred_status]
END as [forwarder_preferred_status],
Sum ([calc_gross_rev]) As [Gr Revenue],
Sum ([charge_weight]) As [Ch Weight]
From
[dbo].[report_bo_awb_revenue_all]
Where
[yr] = 2014
AND [agent_iata_code] not in ('0508634','0514616')
不幸的是，这仍然不起作用，因为您缺少使 SUM() 聚合函数起作用所需的 GROUP BY 子句。由于表达式复杂，我发现通过嵌套查询更容易解决问题，如下所示：
SELECT forwarder_display, [forwarder_preferred_status],
SUM([calc_gross_rev]) As [Gr Revenue],
SUM([charge_weight]) As [Ch Weight]
FROM
(
Select
[forwarder_display],
CASE
WHEN [forwarder_short] IN ('ABL', 'ALK', 'EIF', 'NNR', 'ALI', 'LAF', 'AIT')
THEN 'P2'
WHEN [forwarder_short] IN ('SCHBAX', 'UPS', 'KUE', 'AGI', 'PAN', 'DGF', 'NEC',
'HWE', 'NIS', 'BDP', 'KWE', 'CEVA', 'VIZ', 'FTN', 'UTI')
THEN 'P3'
WHEN [forwarder_short] IN ('OIA','PHO', 'REI')
THEN NULL
ELSE [forwarder_preferred_status]
END as [forwarder_preferred_status],
[calc_gross_rev],
[charge_weight]
From
[dbo].[report_bo_awb_revenue_all]
Where
[yr] = 2014
AND [agent_iata_code] not in ('0508634','0514616')
) t
GROUP BY [forwarder_display], [forwarder_preferred_status]
这应该 运行 并给你想要的结果，除非出现一些错误（很有可能，因为我在答案 window 中输入了所有这些未经测试的内容）。但是，您可能会发现仅通过 copy/pasting 对 GROUP BY 子句的复杂 CASE 表达式就可以获得更好的执行计划，而无需嵌套查询。
最后，像这样的代码——所有这些 3 位和 4 位代码都被硬编码了——要求将这些代码放入 table，以便通过 JOIN + 完成COALESCE，而不是 CASE。这将 MUCH 提高性能和维护。那将允许你写这样的东西：
--Best
SELECT [forwarder_display],
COALESCE(M.status, A.forwarder_preferred_status) As forwarder_preferred_status
Sum ([calc_gross_rev]) As [Gr Revenue],
Sum ([charge_weight]) As [Ch Weight]
FROM [dbo].[report_bo_awb_revenue_all] A
LEFT JOIN forwarder_map M ON m.forwarder_short = A.forwarder_short
WHERE
A.[yr] = 2014
AND [agent_iata_code] not in ('0508634','0514616')
GROUP BY [forwarder_display], COALESCE(M.status, A.forwarder_preferred_status)
最后一点：要使上面的示例起作用，您需要使用空字符串而不是虚构 forwarder_map table 中的 NULL 值。然后，您可以使用 NULLIF() 取回 NULL，如果这确实是您需要的。
你的陈述很难read/edit...你可以使用shorter/simpler方式..像这样：
SELECT [forwarder_display],
CASE
WHEN [forwarder_short] IN ('ABL','ALK','NNR','EIF','ALI','LAF','AIT') THEN 'P2'
WHEN [forwarder_short] IN ('SCHBAX', 'UPS', 'KUE','AGI','PAN'......) THEN 'P3'
WHEN [forwarder_short] IN ('OIA','PHO','REI') THEN null
ELSE [forwarder_preferred_status]
END AS forwarder_preferred_status
...
我敢打赌，这不是第一次有人需要将 ABL 翻译成 P2 来制作报告，也不会是最后一次。正如@Aaron 指出的那样，停止这些可怕的案例陈述并将其表格化。
为了这个答案，我使用了一个 table 变量。这些通常是不可取的，因为数据库引擎不会记录任何关于其中的值及其分布的统计信息，这是制定良好查询计划的关键。
DECLARE
@TRANSLATE table
(
forwarder_short varchar(10) NOT NULL
, forwarder_preferred_status varchar(10) NULL
);
INSERT INTO
@TRANSLATE
(
forwarder_short
, forwarder_preferred_status
)
VALUES
('ABL', 'P2')
, ('ALK', 'P2')
, ('EIF', 'P2')
, ('NNR', 'P2')
, ('ALI', 'P2')
, ('LAF', 'P2')
, ('AIT', 'P2')
, ('SCHBAX', 'P3')
, ('UPS', 'P3')
, ('KUE', 'P3')
, ('AGI', 'P3')
, ('PAN', 'P3')
, ('DGF', 'P3')
, ('NEC', 'P3')
, ('HWE', 'P3')
, ('NIS', 'P3')
, ('BDP', 'P3')
, ('KWE', 'P3')
, ('CEVA', 'P3')
, ('VIZ', 'P3')
, ('FTN', 'P3')
, ('UTI', 'P3')
, ('OIA', NULL)
, ('PHO', NULL)
, ('REI', NULL);
现在我有一个 table、TRANSLATE，它提供了我们已知的 forwarder_short 值到汇总 forwarder_preferred_status 之间的映射。
为了涵盖 forwarder_short 值不在此 table 中的情况，我们需要使用我们的翻译 table 作为 OUTER JOIN。
WITH report_bo_awb_revenue_all AS
(
-- simulating your source data
SELECT
T.forwarder_short
, T.forwarder_preferred_status
, 100 AS calc_gross_rev
, 12 AS charge_weigh
, 2014 AS yr
, '' AS agent_iata_code
, 'fwd' AS forwarder_display
FROM
@TRANSLATE AS T
UNION ALL
SELECT
'UNK'
, 'P90x'
, 9000 AS calc_gross_rev
, 12 AS charge_weigh
, 2014 AS yr
, '' AS agent_iata_code
, 'fwd' AS forwarder_display
)
SELECT
R.forwarder_display
, COALESCE(T.forwarder_preferred_status, R.forwarder_preferred_status) AS forwarder_preferred_status
, SUM(R.calc_gross_rev) AS calc_gross_rev
, SUM(R.charge_weigh) AS charge_weigh
FROM
report_bo_awb_revenue_all AS R
LEFT OUTER JOIN
@TRANSLATE AS T
ON T.forwarder_short = R.forwarder_short
WHERE
R.yr = 2014
AND R.agent_iata_code not in ('0508634','0514616')
GROUP BY
R.forwarder_display
, T.forwarder_preferred_status
, R.forwarder_preferred_status;
现在，所有的业务逻辑都被提炼成了这个
, COALESCE(T.forwarder_preferred_status, R.forwarder_preferred_status) AS forwarder_preferred_status
"Use the value in our translate table if it exists, otherwise, use what's stored on the report_bo_awb_revenue_all table."
SQLFiddle version 这使用临时 table 进行翻译，因为 fiddle 不处理 table 变量，但核心概念仍然存在。