使用此示例数据:
{
"device":"TestDevice",
"id":30,
"flags":[ "New", "Ignore"]
}
我想选择所有没有标志“ Ignore”的数据,我将其与udf一起使用:
SELECT
device, id, flags
FROM input
WHERE udf.ArrayContains(flags, "Ignore) = 0
是否可以在没有用户定义功能的情况下执行此操作?
答案 0 :(得分:1)
这可以解决问题
with cte as
(
select
i.*
from
localInput as i
outer APPLY
getarrayelements(i.flags) ae
where
ae.ArrayValue != 'Ignore'
or
getarraylength(i.flags) = 0
)
select
c.id,
c.device,
c.flags
from
cte c
group by
c.id,
c.device,
c.flags,
System.Timestamp
having
count(*) = getarraylength(c.flags)
or
getarraylength(c.flags) = 0
我用以下示例数据对其进行了测试:
{"device":"TestDevice1","id":1,"flags":[ "New", "Ignore"]}
{"device":"TestDevice2","id":2,"flags":[ "New"]}
{"device":"TestDevice3","id":3,"flags":[ "Ignore"]}
{"device":"TestDevice2","id":4,"flags":[ "Something", "Else"]}
{"device":"TestDevice2","id":5,"flags":[]}