Order Preserving Sorted Table Union

                 Key: PIG-1069
                 URL: https://issues.apache.org/jira/browse/PIG-1069
             Project: Pig
          Issue Type: New Feature
    Affects Versions: 0.6.0
            Reporter: Yan Zhou
            Assignee: Yan Zhou

The output schema will adopt a "schema union" semantics, namely, if an output 
column only appears in one component table, the result rows will have the 
values of the column if the rows are from that component table and null 
otherwise; on the other hand, if an output column appears in multiple component 
tables, the types of the column in all the component tables must be identical. 
Otherwise, an exception will be thrown. The result rows will have the values of 
the column if the rows are from the component tables that have the column 
themselves, or null if otherwise. 

The order preserving sort-unioned results could be further indexed by the 
component tables if the projection contains column(s) named "source_table". If 
so specified, the component table index will be output at the position(s) as 
specified in the projection list. If the underlying table is not a union of 
sorted tables, use of the special column name in projection will cause an 
exception thrown. 

If an attempt is made to create a table of a column named "source_table", an 
excpetion will be thrown as the name is reserved by zebra for the virtual name. 

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