SQL VBA 使用 TRUNC() 函数按错误分组答案

【问题标题】：SQL VBA Group by error with TRUNC() functionSQL VBA 使用 TRUNC() 函数按错误分组
【发布时间】：2020-11-16 07:56:07
【问题描述】：

我正在对 vba excel 进行查询，我对此查询有疑问：

SQLStr = "SELECT DISTINCT (t2.COD_CF) AS CODICE_CLIENTE, (t2.RAG_SOC_CF) AS CLIENTE, (t1.DES_HEAD_DOC) AS COMMESSA_E_DESCRIZIONE, t1.DATA_PREV_FIN_LAV AS GREZZO_CONFERMATO,  MIN(t4.DATA_INIZIO) AS PRIMO_CARICO, MAX(t4.DATA_FINE) AS ULTIMO_CARICO, " & _
"(TRUNC(MAX(t1.DATA_PREV_FIN_LAV)) - TRUNC(t4.DATA_INIZIO) ) - " & _
"((((TRUNC(MAX(t1.DATA_PREV_FIN_LAV),'D'))-(TRUNC(t4.DATA_INIZIO,'D')))/7)*2) - " & _
"(CASE WHEN TO_CHAR(t4.DATA_INIZIO,'DY','nls_date_language=english')='SUN' THEN 1 ELSE 0 END) - " & _
"(CASE WHEN TO_CHAR(MAX(t1.DATA_PREV_FIN_LAV),'DY','nls_date_language=english')='SAT' THEN 1 ELSE 0 END) as GG_DIFFERENZA " & _
"FROM COMM_LAV t1 " & _
"INNER JOIN CF t2 ON t1.COD_CF_INTE = t2.COD_CF " & _
"INNER JOIN COMM_LAV_LNK t3 ON t1.DOC_ID = t3.DOC_ID " & _
"INNER JOIN ORP_EFF_CICLI_ESEC t4 ON t3.LNK_DOC_ID = t4.DOC_ID " & _
"WHERE t1.DATA_PREV_FIN_LAV >= '" & startDate & "' AND t1.DATA_PREV_FIN_LAV <= '" & endDate & "' AND t4.COD_CICLO <> 'LEV01' " & _
"GROUP BY t2.COD_CF, t2.RAG_SOC_CF, t1.DES_HEAD_DOC, t1.DATA_PREV_FIN_LAV " & _
"ORDER BY t1.DES_HEAD_DOC, t1.DATA_PREV_FIN_LAV, MIN(t4.DATA_INIZIO) "

我收到此错误：ORA-00979: not a GROUP BY 表达式只有当我执行 TRUNC() 函数时才会出现此错误，有人可以帮助我并解释这个组

更新我尝试使用 oracle，并像这样编辑查询：

SELECT (t2.COD_CF) AS CODICE_CLIENTE, (t2.RAG_SOC_CF) AS CLIENTE, (t1.DES_HEAD_DOC) AS COMMESSA_E_DESCRIZIONE, t1.DATA_PREV_FIN_LAV AS GREZZO_CONFERMATO,  MIN(t4.DATA_INIZIO) AS PRIMO_CARICO, MAX(t4.DATA_FINE) AS ULTIMO_CARICO,  
(TRUNC(MAX(t1.DATA_PREV_FIN_LAV)) - TRUNC(t4.DATA_INIZIO)) AS DIFF
FROM COMM_LAV t1
INNER JOIN CF t2 ON t1.COD_CF_INTE = t2.COD_CF 
INNER JOIN COMM_LAV_LNK t3 ON t1.DOC_ID = t3.DOC_ID 
INNER JOIN ORP_EFF_CICLI_ESEC t4 ON t3.LNK_DOC_ID = t4.DOC_ID 
WHERE t1.DATA_PREV_FIN_LAV >= '01-OCT-20' AND t1.DATA_PREV_FIN_LAV <= '30-OCT-20' AND t4.COD_CICLO <> 'LEV01' 
GROUP BY t2.COD_CF, t2.RAG_SOC_CF, t1.DES_HEAD_DOC, t1.DATA_PREV_FIN_LAV
ORDER BY t1.DES_HEAD_DOC, t1.DATA_PREV_FIN_LAV, MIN(t4.DATA_INIZIO)

但我总是出现这个错误：ORA-00979: not a GROUP BY expression

为什么 trunc 函数给我这个错误？

【问题讨论】：

您通常 GROUP BY 与您选择的列相同，但那些作为设置函数的参数的列除外。
DISTINCT 不是函数，它是SELECT DISTINCT 的一部分，适用于整个选定的行。
GROUP BY 消除重复行，因此无需执行 SELECT DISTINCT。
我试图不使用 distinct，但我有同样的错误，如果我不把 TRUNC() 函数工作，那为什么？
您应该首先使用 SQL developer 或类似的工具创建您的 SQL 语句。只要它不在那里工作，它也不能从 VBA 工作。仅当它在那里工作时，才将其放入 VBA。 ♦ 使用 ADODB.Parameter 将日期标准传递给 Oracle，请参阅 stackoverflow.com/a/60640185/7599798 ♦ 要检查日期是否为 Sunday，您应该使用 Oracle Weekday 函数

标签： sql vba oracle group-by

【解决方案1】：

GROUP BY 子句需要包含 SELECT 中所有未被 GROUPED 的字段。你省略了这个：

(TRUNC(MAX(t1.DATA_PREV_FIN_LAV)) - TRUNC(t4.DATA_INIZIO)) AS DIFF

即使您在其中有一个聚合函数 (MAX)，总体而言这是一条 TRUNC 语句，因此不是聚合。

我认为逻辑上以下2个语句是相同的（至少只要B

MAX(A) - B == MAX(A-B)

因此，您应该能够将您的语句重写为以下内容，并且仍然得到相同的结果：

MAX(TRUNC(t1.DATA_PREV_FIN_LAV) - TRUNC(t4.DATA_INIZIO)) AS DIFF

【讨论】：

【解决方案2】：

您在GROUP BY 中缺少TRUNC(t4.DATA_INIZIO) 列：

SELECT t2.COD_CF AS CODICE_CLIENTE, t2.RAG_SOC_CF AS CLIENTE,
       t1.DES_HEAD_DOC AS COMMESSA_E_DESCRIZIONE, 
       t1.DATA_PREV_FIN_LAV AS GREZZO_CONFERMATO,
       MIN(t4.DATA_INIZIO) AS PRIMO_CARICO,
       MAX(t4.DATA_FINE) AS ULTIMO_CARICO,  
       (TRUNC(MAX(t1.DATA_PREV_FIN_LAV)) - 
        TRUNC(t4.DATA_INIZIO)
--------^ Not in GROUP BY
       ) AS DIFF
FROM COMM_LAV t1 JOIN
     CF t2 
     ON t1.COD_CF_INTE = t2.COD_CF JOIN
     COMM_LAV_LNK t3
     ON t1.DOC_ID = t3.DOC_ID JOIN
     ORP_EFF_CICLI_ESEC t4
     ON t3.LNK_DOC_ID = t4.DOC_ID JOIN
WHERE t1.DATA_PREV_FIN_LAV >= DATE '2020-10-01' AND 
      t1.DATA_PREV_FIN_LAV <= DATE '2020-10-30' AND
      t4.COD_CICLO <> 'LEV01' 
GROUP BY t2.COD_CF, t2.RAG_SOC_CF, t1.DES_HEAD_DOC, t1.DATA_PREV_FIN_LAV,
         TRUNC(t4.DATA_INIZIO)
---------^ add to GROUP BY
ORDER BY t1.DES_HEAD_DOC, t1.DATA_PREV_FIN_LAV, MIN(t4.DATA_INIZIO);

请注意，Oracle 支持使用 DATE 关键字的日期文字。我强烈建议您使用它。

【讨论】：