[ 
https://issues.apache.org/jira/browse/HIVE-65?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12652192#action_12652192
 ] 

Ashish Thusoo commented on HIVE-65:
-----------------------------------

Actually mysql supports implicit conversions in both directions and that is 
what got me thinking that perhaps we should also be supporting both.

http://dev.mysql.com/doc/refman/5.0/en/type-conversion.html

If we go the polymorphism route we would have to be very careful while adding 
UDFs. The UDF writers will have to really understand how for implicit 
conversion they have to implement functions with different signatures and that 
would just make it that much harder for them to write UDFs that can take 
advantage of implicit conversion. In fact polymorphism in the comparison 
function is the reason why we are hitting this problem in the
first place. 


The problem with going both ways is truly the one that caused us a lot of pain 
initially. However, that does not get solved by restricting the conversion 
possibilities. If for example we have a udf that takes two forms:

a) evaluate(string, string, int)
b) evaluate(int, string, string)

then even if we only allow string to be converted to int and not vice versa we 
have an ambiguous match for

evaluate(1, 'abc', 2)

and we would not know which one to pick. This is what I mean by polymorphism 
causing these issues. The comparison functions are
polymorphic and they are the ones that lead to the problem of ambiguity.


> Implict conversion from integer to long broken for Dynamic Serde tables
> -----------------------------------------------------------------------
>
>                 Key: HIVE-65
>                 URL: https://issues.apache.org/jira/browse/HIVE-65
>             Project: Hadoop Hive
>          Issue Type: Bug
>          Components: Query Processor
>            Reporter: Ashish Thusoo
>            Assignee: Zheng Shao
>
> For a dynamic serde table that has a bigint column, implict conversion from 
> int to bigint seems to be broken. I have not verified this for other tables.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.

Reply via email to