[
https://issues.apache.org/jira/browse/MAPREDUCE-6549?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15008133#comment-15008133
]
Wilfred Spiegelenburg commented on MAPREDUCE-6549:
--------------------------------------------------
[~zxu] The change in MAPREDUCE-6481 is not to blame for the duplicate records
as far as I can tell. It fixed things and now we get to see what is there and
thus we noticed the duplicates. I did not look at the compressed input, and I
do think you are correct. Compressed input uses the same steps and we should
clear the setting in the same way as we did the uncompressed stream. I will try
to generate a compressed stream that is splittable to get a test case. I will
upload a new patch but I will first try to generate the test case before I do
that.
[~cotedm] An EOF will automatically terminate the record there is no need for a
record delimiter at the end of the file. All the test, and comments in the code
show it. The assumption is that the last record before EOF does not need a
record terminator. It is not a new assumption, assuming that an EOF would not
delimit a record would be counter intuitive. Most text files for instance do
not have a newline at the last line.
> multibyte delimiters with LineRecordReader cause duplicate records
> ------------------------------------------------------------------
>
> Key: MAPREDUCE-6549
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-6549
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Components: mrv1, mrv2
> Affects Versions: 2.7.2
> Reporter: Dustin Cote
> Assignee: Wilfred Spiegelenburg
> Attachments: MAPREDUCE-6549-1.patch, MAPREDUCE-6549-2.patch
>
>
> LineRecorderReader currently produces duplicate records under certain
> scenarios such as:
> 1) input string: "abc+++def++ghi++"
> delimiter string: "+++"
> test passes with all sizes of the split
> 2) input string: "abc++def+++ghi++"
> delimiter string: "+++"
> test fails with a split size of 4
> 2) input string: "abc+++def++ghi++"
> delimiter string: "++"
> test fails with a split size of 5
> 3) input string "abc+++defg++hij++"
> delimiter string: "++"
> test fails with a split size of 4
> 4) input string "abc++def+++ghi++"
> delimiter string: "++"
> test fails with a split size of 9
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
