[ 
https://issues.apache.org/jira/browse/PIG-1031?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Olga Natkovich updated PIG-1031:
--------------------------------

    Fix Version/s:     (was: 0.6.0)
                   0.7.0

There is no way to fix this quickly without rethinking of how we deal with 
schemas. Delaying till next release. Workaround to write a custome loader or to 
use type aware one like BinStorage

> PigStorage interpreting chararray/bytearray for a tuple element inside a bag 
> as float or double
> -----------------------------------------------------------------------------------------------
>
>                 Key: PIG-1031
>                 URL: https://issues.apache.org/jira/browse/PIG-1031
>             Project: Pig
>          Issue Type: Bug
>          Components: impl
>    Affects Versions: 0.5.0
>            Reporter: Viraj Bhat
>            Assignee: Daniel Dai
>             Fix For: 0.7.0
>
>
> I have a data stored in a text file as:
> {(4153E765)}
> {(AF533765)}
> I try reading it using PigStorage as:
> {code}
> A = load 'pigstoragebroken.dat' using PigStorage() as 
> (intersectionBag:bag{T:tuple(term:bytearray)});
> dump A;
> {code}
> I get the following results:
> ({(Infinity)})
> ({(AF533765)})
> The problem seems to be with the method: parseFromBytes(byte[] b) in class 
> Utf8StorageConverter. This method uses the TextDataParser (class generated 
> via jjt) to interpret the type of data from content, even though the schema 
> tells it is a bytearray. 
> TextDataParser.jjt  sample code
> {code}
> TOKEN :
> {
> ...
>  < DOUBLENUMBER: (["-","+"])? <FLOATINGPOINT> ( ["e","E"] ([ "-","+"])? 
> <FLOATINGPOINT> )?>
>  < FLOATNUMBER: <DOUBLENUMBER> (["f","F"])? >
> ...
> }
> {code}
> I tried the following options, but it will not work as we need to call 
> bytesToBag(byte[] b) in the Utf8StorageConverter class.
> {code}
> A = load 'pigstoragebroken.dat' using PigStorage() as 
> (intersectionBag:bag{T:tuple(term)});
> A = load 'pigstoragebroken.dat' using PigStorage() as 
> (intersectionBag:bag{T:tuple(term:chararray)});
> {code}
> Viraj

-- 
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