[ https://issues.apache.org/jira/browse/PIG-1074?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Yan Zhou updated PIG-1074: -------------------------- Attachment: PIG-1074.patch To allow for the the column names with "::" on disk to be further used in another COGROUP call before stored in Zebra, multiple "::" should be supported instead of a single one. Therefore, the identifier syntax reads as ( <LETTER> )+ ( <DIGIT> | <LETTER> | <SPECIALCHAR> )* ( <SCOPEOP> ( <LETTER> )+ ( <DIGIT> | <LETTER> | <SPECIALCHAR> )*)* > Zebra store function should allow '::' in column names in output schema > ----------------------------------------------------------------------- > > Key: PIG-1074 > URL: https://issues.apache.org/jira/browse/PIG-1074 > Project: Pig > Issue Type: Bug > Reporter: Pradeep Kamath > Assignee: Yan Zhou > Fix For: 0.6.0, 0.7.0 > > Attachments: PIG-1074.patch, PIG-1074.patch, PIG-1074.patch > > > the following script fails: > {noformat} > a = load '/zebra/singlefile/studenttab10k' using > org.apache.hadoop.zebra.pig.TableLoader() as (name, age, gpa); > b = load '/zebra/singlefile/votertab10k' using > org.apache.hadoop.zebra.pig.TableLoader() as (name, age, registration, > contributions); > c = filter a by age < 20; > d = filter b by age < 20; > store c into > '/user/pig/out//ZebraMultiQuery_30.out.1' using > org.apache.hadoop.zebra.pig.TableStorer(''); > store d into > '/user/pig/out//ZebraMultiQuery_30.out.2' using > org.apache.hadoop.zebra.pig.TableStorer(''); > e = cogroup c by name, d by name; > f = foreach e generate flatten(c), flatten(d); > store f into '/user/pig//ZebraMultiQuery_30.out.3' > using org.apache.hadoop.zebra.pig.TableStorer(''); > {noformat} > Here the schema of f has names like c::name and it looks like zebra storefunc > does not allow '::' in column name > The stack trace is > > ERROR 2997: Unable to recreate exception from backend error: > java.io.IOException: ColumnGroup.Writer constructor failed : Partition > constructor failed :Encountered " ":" ": "" at line 1, column 3. > -- This message is automatically generated by JIRA. - You can reply to this email to add a comment to the issue online.