我在WHERE中有一个带子查询的UPDATE语句来查找重复项。子查询在运行子查询本身时会显示错误，但在UPDATE语句中运行时，不会显示错误，并且DML运行正常（但速度很慢）。
参见表格设置：
CREATE TABLE RAW_table
(
ERROR_LEVEL NUMBER(3),
RAW_DATA_ROW_ID INTEGER,
ATTRIBUTE_1 VARCHAR2(4000 BYTE)
)
;
INSERT INTO RAW_table VALUES (0, 2, '509NTQD9Q868');
INSERT INTO RAW_table VALUES (0, 2, '509NTQD9Q868');
INSERT INTO RAW_table VALUES (0, 2, '509NTQD9Q868');
INSERT INTO RAW_table VALUES (0, 3, '509NTVS9Q863');
INSERT INTO RAW_table VALUES (0, 3, '509NTVS9Q863');
INSERT INTO RAW_table VALUES (0, 3, '509NTVS9Q863');
COMMIT;
有错误的查询是：
SELECT UPPER(ATTRIBUTE_1), rid
FROM ( SELECT UPPER(ATTRIBUTE_1)
, ROWID AS rid
, ROW_NUMBER() OVER ( PARTITION BY UPPER (ATTRIBUTE_1) ORDER BY RAW_DATA_ROW_ID) AS RN
FROM RAW_table
)
WHERE RN > 1;
运行时会显示ORA-00904: "ATTRIBUTE_1": invalid identifier。
但是，在WHERE语句中使用上述查询（从第4行开始）的以下DML工作正常：
set timing on
UPDATE RAW_table
SET ERROR_LEVEL = 4
WHERE (UPPER (ATTRIBUTE_1), ROWID)
IN (SELECT UPPER (ATTRIBUTE_1), rid
FROM (SELECT UPPER (ATTRIBUTE_1), ROWID AS rid
, ROW_NUMBER() OVER ( PARTITION BY UPPER (ATTRIBUTE_1) ORDER BY RAW_DATA_ROW_ID) AS RN
FROM RAW_table
)
WHERE RN > 1
)
;
4 rows updated.
Elapsed: 00:00:00.36
为什么呢？为什么？为什么呢？
我预计UPDATE也会因ORA-00904: "ATTRIBUTE_1": invalid identifier而失败。为什么它不会失败？
然而，真正的问题并不是说UPDATE实际上有效，而是它的工作效果非常慢。
当我更正子查询时，不要触发ORA-00904: "ATTRIBUTE_1": invalid identifier这样的事情：
UPDATE RAW_table
SET ERROR_LEVEL = 4
WHERE (UPPER (ATTRIBUTE_1), ROWID)
IN (SELECT checked_column, rid
FROM (SELECT UPPER (ATTRIBUTE_1) AS checked_column, ROWID AS rid
, ROW_NUMBER() OVER ( PARTITION BY UPPER (ATTRIBUTE_1) ORDER BY RAW_DATA_ROW_ID) AS RN
FROM RAW_table
)
WHERE RN > 1
)
;
查询在11.000行的测试数据集上加速了近400次：
SELECT COUNT(*) FROM RAW_table;
COUNT(*)
----------
11004
1 row selected.
更正了查询：
1005 rows updated.
Elapsed: 00:00:00.28
使用ORA-904查询：
1005 rows updated.
Elapsed: 00:01:48.40
我没耐心等到71.000行结束测试：
SELECT COUNT(*) FROM RAW_table;
COUNT(*)
----------
71475
1 row selected.
Corrected query
11004 rows updated.
Elapsed: 00:00:00.60
Query with ORA-904
30分钟后取消......
使用ORA-904解释查询计划：
UPDATE STATEMENT ALL_ROWS Cost: **2 544 985 615** Bytes: 8 464 752 Cardinality: 4 176
7 UPDATE RAW_TABLE
6 FILTER
1 TABLE ACCESS FULL TABLE RAW_TABLE Cost: 54 Bytes: 169 282 878 Cardinality: 83 514
5 VIEW Cost: 30 486 Bytes: 2 087 850 Cardinality: 83 514
4 WINDOW SORT Cost: 30 486 Bytes: 169 282 878 Cardinality: 83 514
3 FILTER
2 TABLE ACCESS FULL TABLE RAW_TABLE Cost: 54 Bytes: 169 282 878 Cardinality: 83 514
解释纠正查询的计划：
UPDATE STATEMENT ALL_ROWS Cost: **36 637** Bytes: 3 374 235 Cardinality: 835
7 UPDATE RAW_TABLE
6 HASH JOIN RIGHT SEMI Cost: 36 637 Bytes: 3 374 235 Cardinality: 835
4 VIEW VIEW SYS.VW_NSO_1 Cost: 30 486 Bytes: 168 197 196 Cardinality: 83 514
3 VIEW Cost: 30 486 Bytes: 169 282 878 Cardinality: 83 514
2 WINDOW SORT Cost: 30 486 Bytes: 169 282 878 Cardinality: 83 514
1 TABLE ACCESS FULL TABLE RAW_TABLE Cost: 54 Bytes: 169 282 878 Cardinality: 83 514
5 TABLE ACCESS FULL TABLE RAW_TABLE Cost: 54 Bytes: 169 282 878 Cardinality: 83 514
分析表后，成本计划是一样的。 使用ORA-904解释查询计划：
UPDATE STATEMENT ALL_ROWS Cost: **29 381 690** Bytes: 38 Cardinality: 2
7 UPDATE RAW_TABLE
6 FILTER
1 TABLE ACCESS FULL TABLE RAW_TABLE Cost: 54 Bytes: 1 358 025 Cardinality: 71 475
5 VIEW Cost: 427 Bytes: 1 786 875 Cardinality: 71 475
4 WINDOW SORT Cost: 427 Bytes: 1 358 025 Cardinality: 71 475
3 FILTER
2 TABLE ACCESS FULL TABLE RAW_TABLE Cost: 54 Bytes: 1 358 025 Cardinality: 71 475
解释纠正查询的计划：
UPDATE STATEMENT ALL_ROWS Cost: **3 123** Bytes: 1 453 595 Cardinality: 715
7 UPDATE RAW_TABLE
6 HASH JOIN SEMI Cost: 3 123 Bytes: 1 453 595 Cardinality: 715
5 VIEW VIEW SYS.VW_NSO_1 Cost: 427 Bytes: 143 950 650 Cardinality: 71 475
4 VIEW Cost: 427 Bytes: 144 879 825 Cardinality: 71 475
3 WINDOW SORT Cost: 427 Bytes: 1 358 025 Cardinality: 71 475
2 TABLE ACCESS FULL TABLE RAW_TABLE Cost: 54 Bytes: 1 358 025 Cardinality: 71 475
1 TABLE ACCESS FULL TABLE RAW_TABLE Cost: 54 Bytes: 1 358 025 Cardinality: 71 475
解释计划成本说明了一切，但为什么会有这么多不同？
我刚刚在计算了桌子上的统计数据后再次触发了71.000行测试，但它已经运行了几分钟......
这一切都在Oracle Database 12c企业版12.1.0.2.0版 - 64位上。
答案 0 :(得分：3)
您的SELECT失败，因为子查询中没有名为ATTRIBUTE_1的列。您需要指定名称：
SELECT UPPER(ATTRIBUTE_1), rid
FROM ( SELECT UPPER(ATTRIBUTE_1) as ATTRIBUTE_1,
ROWID AS rid,
ROW_NUMBER() OVER (PARTITION BY UPPER(ATTRIBUTE_1) ORDER BY RAW_DATA_ROW_ID) AS RN
FROM RAW_table
)
WHERE RN > 1;
UPDATE不会生成错误，因为它从外部查询中提取值：
UPDATE RAW_table
-------^
| SET ERROR_LEVEL = 4
| WHERE (UPPER (ATTRIBUTE_1), ROWID) IN
| (SELECT checked_column, rid
| FROM (SELECT UPPER(ATTRIBUTE_1) AS checked_column, ROWID AS rid,
------------------------------^ This is interpreted as RAW_table.ATTRIBUTE_1
ROW_NUMBER() OVER (PARTITION BY UPPER(ATTRIBUTE_1) ORDER BY RAW_DATA_ROW_ID) AS RN
FROM RAW_table
)
WHERE RN > 1
)
这种相关性可能不是您想要的，也是我建议列名始终是合格的（即包含表别名）的一个原因。
答案 1 :(得分：2)
这就是为什么别名确实非常有用。
在查询中
UPDATE RAW_table
SET ERROR_LEVEL = 4
WHERE (UPPER (ATTRIBUTE_1), ROWID)
IN (SELECT UPPER (ATTRIBUTE_1), rid
FROM (SELECT UPPER (ATTRIBUTE_1), ROWID AS rid
, ROW_NUMBER() OVER ( PARTITION BY UPPER (ATTRIBUTE_1)
ORDER BY RAW_DATA_ROW_ID) AS RN
FROM RAW_table
)
WHERE RN > 1
)
SELECT UPPER (ATTRIBUTE_1)有效，因为它可以解析为对您正在更新的表的引用而不是FROM中的表。对于别名，该查询等同于
UPDATE RAW_table dest
SET dest.ERROR_LEVEL = 4
WHERE (UPPER (dest.ATTRIBUTE_1), ROWID)
IN (SELECT UPPER (dest.ATTRIBUTE_1), src.rid
FROM (SELECT UPPER (rt.ATTRIBUTE_1), rt.ROWID AS rid
, ROW_NUMBER() OVER ( PARTITION BY UPPER (rt.ATTRIBUTE_1)
ORDER BY rt.RAW_DATA_ROW_ID) AS RN
FROM RAW_table rt
) src
WHERE src.rid > 1
)
当然，如果您以这种方式编写，则会立即明确您引用dest.attribute_1而不是src.attribute_1。这（以及许多其他原因）是为什么给列添加别名是个好主意 - 它清楚地说明了您想要引用哪个对象，并在预期引用无效时抛出错误，而不是将其解析为您没有的内容打算。
答案 2 :(得分：0)
Patient::where('id', $patientId)->update(['doctor_id' => $result->doctor_id]);
答案 3 :(得分：0)
也许这些版本更快（至少它们更紧凑）：
UPDATE RAW_table
SET ERROR_LEVEL = 4
WHERE ROWID <>ALL (SELECT MIN(ROWID) FROM RAW_table GROUP BY UPPER(ATTRIBUTE_1));
UPDATE RAW_table
SET ERROR_LEVEL = 4
WHERE ROWID <>ALL (SELECT FIRST_VALUE(ROWID) OVER (PARTITION BY UPPER(ATTRIBUTE_1) ORDER BY RAW_DATA_ROW_ID) FROM RAW_table);
注意，<>ALL相当于NOT IN - 我个人倾向于使用<>ALL。