[ 
https://issues.apache.org/jira/browse/PIG-592?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12738472#action_12738472
 ] 

Daniel Dai commented on PIG-592:
--------------------------------

Also the following script produce the wrong schema:

a = load 'a';
b = load 'b';
c = join a by $0, b by $0;
describe c;

c: {bytearray,bytearray}

The correct behavior should be: If any of the input schema is unkown, the 
output schema is also unkown. 

> schema inferred incorrectly
> ---------------------------
>
>                 Key: PIG-592
>                 URL: https://issues.apache.org/jira/browse/PIG-592
>             Project: Pig
>          Issue Type: Bug
>    Affects Versions: 0.2.0
>            Reporter: Christopher Olston
>
> A simple pig script, that never introduces any schema information:
> A = load 'foo';
> B = foreach (group A by $8) generate group, COUNT($1);
> C = load 'bar';       // ('bar' has two columns)
> D = join B by $0, C by $0;
> E = foreach D generate $0, $1, $3;
> Fails, complaining that $3 does not exist:
> java.io.IOException: Out of bound access. Trying to access non-existent 
> column: 3. Schema {B::group: bytearray,long,bytearray} has 3 column(s).
> Apparently Pig gets confused, and thinks it knows the schema for C (a single 
> bytearray column).

-- 
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