gt; >> > I'm a little of doubt about the spark's behavior. But from my sides,
> >> seems Spark actually use string type which is different from your
> >> investigation.
> >> > I try with the following sql in spark-cli:
> >> > `
> >> > select array() + 1
>
`ARRAY`/ `INT ARRAY` is for declaring data type in DDL like `create table
t1(a INT ARRAY)`.
Best regards,
Yuxia
- 原始邮件 -
发件人: "eric xiao"
收件人: "dev"
发送时间: 星期六, 2022年 10 月 29日 上午 7:50:22
主题: Re: [DISCUSS] FLINK-20578 Cannot create empty array using ARRAY[]
Upon f
: differing types in '(array() + 1)' (array and int).; line
>> 1 pos 7;
>> > 'Project [unresolvedalias((array() + 1), None)]
>> > +- OneRowRelation
>> > `
>> >
>> >
>> > Seems it's hard to decide which data type Flink should use. I'm
>>
;
> > Seems it's hard to decide which data type Flink should use. I'm
> insterested in the reason why you would like to use Integer type.
> > I haven't cheked whether the sql stardard specifies it. But from my
> side, I prefer to follow Hive/Spark.
> >
> > BTW: the quer
: the query `SELECT COALESCE(1, cast(ARRAY() as int))` will fail in Hive and
Spark.
Best regards,
Yuxia
- 原始邮件 -
发件人: "eric xiao"
收件人: "dev"
发送时间: 星期四, 2022年 10 月 27日 下午 9:13:51
主题: [DISCUSS] FLINK-20578 Cannot create empty array using ARRAY[]
Hi,
I would like t
.
BTW: the query `SELECT COALESCE(1, cast(ARRAY() as int))` will fail in Hive and
Spark.
Best regards,
Yuxia
- 原始邮件 -
发件人: "eric xiao"
收件人: "dev"
发送时间: 星期四, 2022年 10 月 27日 下午 9:13:51
主题: [DISCUSS] FLINK-20578 Cannot create empty array using ARRAY[]
Hi,
I w
Hi,
I would like to propose a solution to this JIRA issue. I looked at the
comments and there was some guidance around where in the code we should
update to allow for this behaviour. But I believe there are still two
questions that remain open:
1. Is this expected behaviour (i.e. users should