hive-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Matt Traynham (JIRA)" <j...@apache.org>
Subject [jira] [Updated] (HIVE-10366) Multiple Date Functions inside Named Struct return the wrong values
Date Thu, 16 Apr 2015 18:03:59 GMT

     [ https://issues.apache.org/jira/browse/HIVE-10366?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Matt Traynham updated HIVE-10366:
---------------------------------
    Description: 
Using multiple date functions inside a named struct results in incorrect values.  The first
date function is correct.  All subsequent date functions receive the value of the first date
function, unless the function has a different valued argument.

As shown below, the 3rd query returns an incorrect value for `month(datetime)`.  A not so
pleasant workaround is to substr datetime resulting in a different value (4th query).

> DESCRIBE my_datamart;
OK
datetime            	string

> SELECT datetime FROM my_datamart LIMIT 1;
OK
2015-04-14T00:00:00

> SELECT named_struct('dayofmonth', day(datetime), 'monthofyear', month(datetime)), day(datetime),
month(datetime) FROM my_datamart LIMIT 1;
OK
{"dayofmonth":14,"monthofyear":14} 14	4

> SELECT named_struct('dayofmonth', day(datetime), 'monthofyear', month(substr(datetime,
0, 10))), day(datetime), month(datetime) FROM my_datamart LIMIT 1;
OK
{"dayofmonth":14,"monthofyear":4} 14	4

I havn't tried all the other date functions, but this is happens with  weekofyear as well.
 Looks like a caching issue...

  was:
Using multiple date functions inside a named struct results in incorrect values.  The first
date function is correct.  All subsequent date functions receive the value of the first date
function, unless the function has a different valued argument.

As shown below, the 3rd query returns an incorrect value for `month(datetime)`.  A not so
pleasant workaround is to substr datetime resulting in a different value (4th query).

> DESCRIBE my_datamart;
OK
datetime            	string

> SELECT datetime FROM my_datamart LIMIT 1;
OK
2015-04-14T00:00:00

> SELECT named_struct('dayofmonth', day(datetime), 'monthofyear', month(datetime)), day(datetime),
month(datetime) FROM my_datamart LIMIT 1;
OK
{"dayofmonth":14,"monthofyear":14}	14	4

> SELECT named_struct('dayofmonth', day(datetime), 'monthofyear', month(substr(datetime,
0, 10))), day(datetime), month(datetime) FROM my_datamart LIMIT 1;
OK
{"dayofmonth":14,"monthofyear":4}	14	4

I havn't tried all the other date functions, but this is happens with  weekofyear as well.
 Looks like a caching issue...


> Multiple Date Functions inside Named Struct return the wrong values
> -------------------------------------------------------------------
>
>                 Key: HIVE-10366
>                 URL: https://issues.apache.org/jira/browse/HIVE-10366
>             Project: Hive
>          Issue Type: Bug
>          Components: Hive
>    Affects Versions: 0.14.0
>            Reporter: Matt Traynham
>
> Using multiple date functions inside a named struct results in incorrect values.  The
first date function is correct.  All subsequent date functions receive the value of the first
date function, unless the function has a different valued argument.
> As shown below, the 3rd query returns an incorrect value for `month(datetime)`.  A not
so pleasant workaround is to substr datetime resulting in a different value (4th query).
> > DESCRIBE my_datamart;
> OK
> datetime            	string
> > SELECT datetime FROM my_datamart LIMIT 1;
> OK
> 2015-04-14T00:00:00
> > SELECT named_struct('dayofmonth', day(datetime), 'monthofyear', month(datetime)),
day(datetime), month(datetime) FROM my_datamart LIMIT 1;
> OK
> {"dayofmonth":14,"monthofyear":14} 14	4
> > SELECT named_struct('dayofmonth', day(datetime), 'monthofyear', month(substr(datetime,
0, 10))), day(datetime), month(datetime) FROM my_datamart LIMIT 1;
> OK
> {"dayofmonth":14,"monthofyear":4} 14	4
> I havn't tried all the other date functions, but this is happens with  weekofyear as
well.  Looks like a caching issue...



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message