[
https://issues.apache.org/jira/browse/SQOOP-1125?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13702322#comment-13702322
]
Kyrill Alyoshin commented on SQOOP-1125:
----------------------------------------
Basically, you likely want to always ROUND_UP in
protected BigDecimal tryDivide(BigDecimal numerator, BigDecimal denominator) {
try {
return numerator.divide(denominator);
} catch (ArithmeticException ae) {
return numerator.divide(denominator, BigDecimal.ROUND_HALF_UP);
}
}
instead of ROUND_HALF_UP which gives a zero when getting a result of any
division less that 0.5.
> Out of memory errors when number of records to import < 0.5 * splitSize
> -----------------------------------------------------------------------
>
> Key: SQOOP-1125
> URL: https://issues.apache.org/jira/browse/SQOOP-1125
> Project: Sqoop
> Issue Type: Bug
> Affects Versions: 1.4.3
> Reporter: Dave Kincaid
> Priority: Critical
>
> We are getting out of memory errors during import if the number of records to
> import is less than 0.5*splitSize (and is nonterminating decimal).
> For example, if the numSplits = 3, minVal = 100, maxVal = 101 then in
> BigDecimalSplitter.split() an extraordinary number of tiny values will be
> added to the splits List and run out of memory eventually.
> I also noticed that there are no tests for BigDecimalSplitter.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
