[ https://issues.apache.org/jira/browse/PIG-1787?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Daniel Dai updated PIG-1787: ---------------------------- Attachment: PIG-1787-2.patch PIG-1787-2.patch fix unit test failures. > Error in logical plan generated > ------------------------------- > > Key: PIG-1787 > URL: https://issues.apache.org/jira/browse/PIG-1787 > Project: Pig > Issue Type: Bug > Affects Versions: 0.8.0 > Reporter: Anitha Raju > Assignee: Daniel Dai > Attachments: PIG-1787-1.patch, PIG-1787-2.patch > > > Here is a sample pig script: > set default_parallel 2 > ALLDATA = load 'sample.txt' using PigStorage() as (id, spaceid, type, pcid); > C1 = filter ALLDATA by (type == 'p' and > (spaceid == '1196250013' > or spaceid == '1196250024' > or spaceid == '1196250011')); > C2 = group C1 by pcid; > C3 = foreach C2 generate flatten(group) as (pc_id), COUNT(C1) as tot; > C4 = order C3 by tot desc; > C5 = limit C4 3; > C6 = join C5 by pc_id, C1 by pcid; > dump C6; > sample.txt: > 1 1196250013 p 1234 > 2 1196250024 p 2314 > 3 1196250011 t 1111 > 4 1111111111 p 1231 > 5 1196250013 p 1254 > 6 1196250024 p 9007 > This fails with the error > java.io.IOException: Type mismatch in key from map: expected > org.apache.pig.impl.io.NullableLongWritable, recieved > org.apache.pig.impl.io.NullableBytesWritable > when both pc_id and pcid are of type bytearray. > The script seems to work when > a) replicated join is substituted in the place of the regular join > b) pcid is cast to long in the loader > c) doing a dump of any statement before C6 > d) setting default_parallel to 1 or removing it. > > One possible cause seems to be with the logical plan generation during the > projection operation in C4 as can be observed from the describe statement. -- This message is automatically generated by JIRA. - You can reply to this email to add a comment to the issue online.