如何将字符串列拆分为包含 NULL 的行？答案

【问题标题】：How to split a string column into rows including NULLs?如何将字符串列拆分为包含 NULL 的行？
【发布时间】：2021-10-24 03:54:15
【问题描述】：

我想将一个字符串列拆分为包含 NULL 值的行。

我使用以下，但是，它不返回 NULL 行。有没有其他方法可以在不消除 NULL 行的情况下实现这一点？

SELECT TRIM(REGEXP_REPLACE(s.value, '[{}_]', ' ')) AS extracted
FROM t1, LATERAL SPLIT_TO_TABLE(val, ',') s

以下是示例数据：

WITH t1 AS (
SELECT 'A' AS id, '{Other}' AS val UNION ALL
SELECT 'A' AS id, NULL AS val UNION ALL
SELECT 'A' AS id, '{Other Questions,Missing Document}' AS val UNION ALL
SELECT 'A' AS id, NULL AS val )

期望的输出：

+----+------------------+
| id |    extracted     |
+----+------------------+
| A  | Other            |
| A  |                  |
| A  | Other Questions  |
| A  | Missing Document |
| A  |                  |
+----+------------------+

【问题讨论】：

标签： sql snowflake-cloud-data-platform flatten

【解决方案1】：

使用LEFT JOIN：

SELECT TRIM(REGEXP_REPLACE(s.value, '[{}_]', ' ')) AS extracted
FROM t1 LEFT JOIN LATERAL
     SPLIT_TO_TABLE(t1.val, ',') s
     ON 1=1;

【讨论】：

【解决方案2】：

您可以在横向拆分表格上放置一个 IFF：

WITH t1 AS (
SELECT 'A' AS id, '{Other}' AS val UNION ALL
SELECT 'A' AS id, NULL AS val UNION ALL
SELECT 'A' AS id, '{Other Questions,Missing Document}' AS val UNION ALL
SELECT 'A' AS id, NULL AS val )


SELECT TRIM(REGEXP_REPLACE(s.value, '[{}_]', ' ')) AS extracted
FROM t1, LATERAL SPLIT_TO_TABLE(iff(val is null, '', val), ',') s

;

【讨论】：