Question

我要用TAB_B中的随机记录（5％）填充一个空表（TAB_A），并与其他3个表（C1，C2，C3）连接。

现在，如果我运行以下查询，一切运行良好：

INSERT INTO TAB_A (field1,field2)
    SELECT TAB_B.ID_TASS, TAB_B.ID_SEZ
    FROM (SELECT TAB_B.*, row_number() 
            OVER (ORDER BY dbms_random.VALUE) r 
            FROM TAB_B) 
    WHERE r < (SELECT COUNT(*) FROM TAB_B)*0.05

问题是我还需要填写TAB_A.field3，其值在TAB_C3中。所以我重写了我之前的查询：

INSERT INTO TAB_A (field1,field2,field3)
    SELECT B.ID_TASS, B.ID_SEZ, C3.Z_ID
    FROM (SELECT B.*, row_number() 
            OVER (ORDER BY dbms_random.VALUE) r 
            FROM TAB_B B
                JOIN TAB_C1 C1 on C1.X_ID = B.X_ID
                JOIN TAB_C2 C2 on C2.Y_ID = C1.Y_ID
                JOIN TAB_C3 C3 on C3.Z_ID = C2.Z_ID
            WHERE C3.Z_ID = 9) 
    WHERE r < (SELECT COUNT(*) FROM B)*0.05

但发生错误：ORA-00904：“C3.Z_ID”无效标识符。

我认为问题可能出现在“ SELECT B.ID_TASS，B.ID_SEZ，C3.Z_ID ”这句话中，我在“ SELECT TAB_B.ID_TASS，TAB_B.ID_SEZ”中进行了更改，TAB_C3.Z_ID “......但没有成功：再次发生同样的错误。

顺便说一句，以下查询完全有效：

select count(*)
    from TAB_B B
        JOIN TAB_C1 C1 on C1.X_ID = B.X_ID
        JOIN TAB_C2 C2 on C2.Y_ID = C1.Y_ID
        JOIN TAB_C3 C3 on C3.Z_ID = C2.Z_ID
    WHERE C3.Z_ID = 9

有人能帮助我吗？

Answer 1

尝试

...

SELECT B.ID_TASS，B.ID_SEZ， C3_Z_ID

FROM（SELECT B. *， C3.Z_ID AS C3_Z_ID， row_number（） ...

Answer 2

虽然我没有测试

，但以下情况应该有效

INSERT INTO TAB_A (field1,field2,field3)
>     SELECT intTab.ID_TASS, intTab.ID_SEZ, intTab.Z_ID
>     FROM (SELECT B.*, c3.z_id,row_number() 
>             OVER (ORDER BY dbms_random.VALUE) r 
>             FROM TAB_B B
>                 JOIN TAB_C1 C1 on C1.X_ID = B.X_ID
>                 JOIN TAB_C2 C2 on C2.Y_ID = C1.Y_ID
>                 JOIN TAB_C3 C3 on C3.Z_ID = C2.Z_ID
>             WHERE C3.Z_ID = 9) intTab
>     WHERE intTab.r < (SELECT COUNT(*) FROM B)*0.05

SQL - 从连接表中提取随机样本

2 个答案: