相当于Spark SQL中的LISTAGG()

Question

我正在Spark SQL中重写Redshift SQL。由于Spark SQL中不支持LISTAGG()，是否有等效的函数或解决方案来实现这一点?

红移SQL:

SELECT
dp_info_id,
dp_type,
CASE
WHEN COALESCE(type,'-1') = 'Primary Name' 
THEN LISTAGG(DISTINCT fir_name,'|') WITHIN GROUP (ORDER BY dp_info_id) 
ELSE NULL
END AS primary_first_name,
FROM
dp_info c
GROUP BY
dp_info,
type,
dp_type

Answer 1

要从组中获得所有值的数组我猜你应该使用collect_set (https://docs.databricks.com/sql/language-manual/functions/collect_list.html)