return value is wrong

[周胜为 <86...@qq.com>]

I develop a UDA function as follow:
input:string
Output:string
I want to caculate every name' md5 code and then caculate the md5 code with XOR. I will get one value which is all md5 code caculate.


md5 code is saved as unsigned char [16], as I  know StringVal has a member called ptr(uint8_t *) I set the md5 code(caculate result) as StringVal' ptr.


Every time I get null value,that is why?








------------------ 原始邮件 ------------------
发件人: "skyyws"<sk...@163.com>;
发送时间: 2018年6月6日(星期三) 上午10:08
收件人: "dev@impala.apache.org"<de...@impala.apache.org>;

主题: Re: A question about AnalysisException



Thanks for your reply, and I knew that both column number and alias worked like this:
------
select datediff(day,now()) from test_table where day>='2018-06-01' group by 1
select datediff(day,now()) d from test_table where day>='2018-06-01' group by d
------
I just wonder why built-in function in where clause instead of constants would resulting in this exception. It's impala syntax ?
On 06/5/2018 21:54，Mike Labman<mi...@gwsolutions.com> wrote：
Just use column number in your group by.

select datediff(day,now()) from test_table where day>='2018-06-01' group by 1

-----Original Message-----
From: skyyws [mailto:skyyws@163.com]
Sent: Tuesday, June 05, 2018 9:49 AM
To: dev
Subject: Re: A question about AnalysisException

Here is the corret result of the sql below:
select datediff(day,now()) from test_table where day>='2018-06-01' group by datediff(day,now());

| datediff(day, now()) |
+----------------------+
| -4                   |
| 0                    |
| -3                   |
| -1                   |
| -2                   |




2018-06-05
skyyws



发件人：skyyws <sk...@163.com>
发送时间：2018-06-05 21:44
主题：A question about AnalysisException
收件人："dev@impala.apache.org"<de...@impala.apache.org>
抄送：

Hello all,
Recently, I found a probelm when I used impala to do ad-hoc analysis. When I executed the sql below:
select datediff(day,now()) from test_table where day>=(now() - interval 5 days) group by datediff(day,now());
I got an exception like this:
-----------------------------------------------------------------------------------------------------------------------------
Status: AnalysisException: select list expression not produced by aggregation output (missing from GROUP BY clause?): datediff(day, TIMESTAMP '2018-06-05 21:24:28.403393000')
-----------------------------------------------------------------------------------------------------------------------------
and if I execute this sql:
select datediff(day,now()) from test_table where day>='2018-06-01' group by datediff(day,now());
I got the correct result like this:

This situation happend both on 2.10.0 and 3.0.0 version.
I'm not sure it's a bug or it's just designed like this, anyone who can give me some advice? Thanks.
(test_table is stored as parquet, and day is the partition column, string type.)
--------------------------------------------------------------------------------
2018-06-05
skyyws

________________________________

Confidentiality Note: This e-mail, and any attachment to it, contains privileged and confidential information intended only for the use of the individual(s) or entity named on the e-mail. If the reader of this e-mail is not the intended recipient, or the employee or agent responsible for delivering it to the intended recipient, you are hereby notified that reading it is strictly prohibited. If you have received this e-mail in error, please immediately return it to the sender and delete it from your system. Thank you