字符串转义不一致 Snowflake 与 dbt

Question

我正在清理一个类似数组的字符串，以便它是具有正确引用的正确数组类型。

输入示例：

['Home & Kitchen / Kitchen & Dining / Dining & Entertaining / Glassware & Drinkware'  'Home & Kitchen / Kitchen & Dining / Dining & Entertaining / Novelty / Drinkware / Coffee Mugs']

当前代码：

array_remove(array_remove(split(replace(replace(categories, '['), ']'), '\''), ''::variant), '\n '::variant)

输出：

[
  "Home & Kitchen / Kitchen & Dining / Dining & Entertaining / Glassware & Drinkware",
  "Home & Kitchen / Kitchen & Dining / Dining & Entertaining / Novelty / Drinkware / Coffee Mugs"
]

尽管字符串转义，但当我应用于 dbt 中的模型时，该解决方案并不适用。我收到下一列“意外”的语法错误。

我发现这个问题与外部替换处理单引号直接相关，即

'\"'

已经旨在安全地处理替换单引号，但显然不在 dbt 中。

欢迎任何想法/建议！

Answer 1

以下方法应适用于您提供的数据：

select 
  ['Home & Kitchen / Kitchen & Dining / Dining & Entertaining / Glassware & Drinkware','Home & Kitchen / Kitchen & Dining / Dining & Entertaining / Novelty / Drinkware / Coffee Mugs']::variant 
    as my_variant,
  array_distinct(strtok_to_array(array_to_string(my_variant, '/'), '/'));

返回：

[   "Home & Kitchen ",   " Kitchen & Dining ",   " Dining & Entertaining ",   " Glassware & Drinkware",   " Novelty ",   " Drinkware ",   " Coffee Mugs" ]