Question

我正在尝试使用xmlagg将行转换为逗号分隔列表。我遇到的麻烦是col2中的重复值。我需要为col2中的每个重复值添加一个新行。

这是源数据：

+-----------------+-----------------+------------+
|      Col1       |        Col2     |      Col3  |
+-----------------+-----------------+------------+
| VSPARK 27002026 | account_id      | account_id |
| VSPARK 27002026 | account_name    | Kontod 6-7 |
| VSPARK 27002026 | account_name    | Kontod 1-3 |
| VSPARK 27002026 | account_name    | Kontod 4-6 |
| VSPARK 27002026 | cash_type       | CASH_TYP   |
| VSPARK 27002026 | Currency        | CURRENCY   |
| VSPARK 27002026 | cust_type       | CUST_TYP   |
| VSPARK 27002026 | Residency       | RESIDENCY  |
| VSPARK 27002026 | transaction_amt | AMT        |
+-----------------+-----------------+------------+

这是预期的输出：

+-----------------+--------------------------------------------------------------------------------------+-----------------------------------------------------------------------+
|      Col1       |                                         Col2                                         |                                 Col3                                  |
+-----------------+--------------------------------------------------------------------------------------+-----------------------------------------------------------------------+
| VSPARK 27002026 | account_name, account_id, cash_type, cust_type, Residency, transaction_amt, Currency |  Kontod 6-7, account_id, CASH_TYP, CUST_TYP, RESIDENCY, AMT, CURRENCY |
| VSPARK 27002026 | account_name, account_id, cash_type, cust_type, Residency, transaction_amt, Currency |  Kontod 1-3, account_id, CASH_TYP, CUST_TYP, RESIDENCY, AMT, CURRENCY |
| VSPARK 27002026 | account_name, account_id, cash_type, cust_type, Residency, transaction_amt, Currency |  Kontod 4-6, account_id, CASH_TYP, CUST_TYP, RESIDENCY, AMT, CURRENCY |
+-----------------+--------------------------------------------------------------------------------------+-----------------------------------------------------------------------+

col2中的每个重复值都应在最终结果中创建一个额外的行。不幸的是我不知道如何做到这一点。

欢迎任何帮助！

EDIT1：

这是我目前的SQL：

SELECT report_name,
   TRIM(TRAILING ','
        FROM (XMLAGG(TRIM(insert_into_fields)|| ','
                     ORDER BY report_name) (VARCHAR(10000)))),
   TRIM(TRAILING ','
        FROM (XMLAGG(TRIM(select_from_fields)|| ','
                     ORDER BY report_name) (VARCHAR(10000))))
FROM
(SELECT report_name,
        field_name insert_into_fields,
        CASE
            WHEN account_mapping_formula_id IS NOT NULL THEN mapping_name
            WHEN formula IS NULL
                 AND account_mapping_Formula_id IS NULL THEN source_field_name
            ELSE formula
        END select_from_fields
 FROM mapping rf
 LEFT JOIN report r ON rf.report_id = r.report_id
 LEFT JOIN report_field e ON rf.target_column_id = e.field_id
 LEFT JOIN source_fields c ON rf.source_field_id = c.source_field_id
 LEFT JOIN account_mapping d ON rf.account_mapping_formula_id = d.account_mapping_id
 WHERE rf.report_id = 2) asd
GROUP BY report_name;

这是当前的输出：

+-----------------+-----------------------------------------------------------------------------------------------------------------+-----------------------------------------------------------------------------------------------+
|      Col1       |                                                      Col2                                                       |                                             Col3                                              |
+-----------------+-----------------------------------------------------------------------------------------------------------------+-----------------------------------------------------------------------------------------------+
| VSPARK 27002026 | account_name, account_name, account_name, account_id, cash_type,cust_type, Residency, transaction_amt, Currency |  Kontod 6-7, Kontod 1-3, Kontod 4-6, account_id, CASH_TYP, CUST_TYP, RESIDENCY, AMT, CURRENCY |
+-----------------+-----------------------------------------------------------------------------------------------------------------+-----------------------------------------------------------------------------------------------+

Answer 1

哎呀，这不是一个用户友好的数据模型:)

您需要复制每个account_name的其他行。一种解决方案是加入(Col1,Col3)的不同列表，然后将Col3添加到最终GROUP BY，类似这样的

WITH cte AS 
 (
   SELECT * 
   FROM asd -- this is your current Derived Table
 )
SELECT t2.report_name,
   'account_name' || ','||
   Trim(Trailing ','
        FROM (XmlAgg(Trim(t1.insert_into_fields)|| ','
                     -- Either sort by a unique column or no sort at all
                     ORDER BY t1.insert_into_fields) (VARCHAR(10000)))),
   t2.select_from_fields || ','||
   Trim(Trailing ','
        FROM (XmlAgg(Trim(t1.col3)|| ','
                     -- Either sort by a unique column or no sort at all
                     ORDER BY t1.insert_into_fields) (VARCHAR(10000))))
FROM cte AS t1
JOIN 
 ( -- list of report/account combinations
   SELECT DISTINCT report_name, select_from_fields
   FROM cte 
   WHERE insert_into_fields = 'account_name'
 ) AS t2
ON t1.report_name = t2.report_name
WHERE t1.insert_into_fields <> 'account_name'
GROUP BY t2.report_name, t2.select_from_fields

将teradata重复行合并为一行

1 个答案: