You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@sqoop.apache.org by Zoltán Tóth <sw...@gmail.com> on 2017/08/29 11:20:54 UTC
Re: Review Request 59833: SQLServerDatatypeImportDelimitedFileTest
can fail in some environments
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/59833/#review184036
-----------------------------------------------------------
Ship it!
Hey Szabolcs,
I ran the unit and the third party tests successfully with your patch.
Thanks for your contribution
Cheers, Zoli
- Zoltán Tóth
On June 6, 2017, 10:29 a.m., Szabolcs Vasas wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/59833/
> -----------------------------------------------------------
>
> (Updated June 6, 2017, 10:29 a.m.)
>
>
> Review request for Sqoop.
>
>
> Bugs: sqoop-3195
> https://issues.apache.org/jira/browse/sqoop-3195
>
>
> Repository: sqoop-trunk
>
>
> Description
> -------
>
> I tried to execute the SQLServer third party tests with two different SQL Server versions. With SQL Server 2014 Express edition all the tests were successful however with SQL Server 2017 Developer edition I got the following error:
> [junit] Test org.apache.sqoop.manager.sqlserver.SQLServerDatatypeImportDelimitedFileTest FAILED
> Failure for following Test Data :
> FLOAT
> SCALE : null
> PREC : null
> TO_INSERT : 1.7976931348623157
> DB_READBACK : 1.7976931348623155
> HDFS_READBACK : 1.7976931348623155
> NEG_POS_FLAG : POS
> OFFSET : 8
> -----------------------------------------------------------
> Exception details :
> expected a different string expected:<1.797693134862315[5]> but was:<1.797693134862315[7]>
> By looking at the test case I have found that it inserts 1.7976931348623157 into the database but it expects 1.7976931348623155 (the last digit is 5 instead of 7) probably because float is an approximate numeric data types on MSSQL and on earlier versions this is how it worked but.
> I suggest using a less precise float number in this test case to avoid flakyness.
>
>
> Diffs
> -----
>
> testdata/DatatypeTestData-import-lite.txt a4b5c75
>
>
> Diff: https://reviews.apache.org/r/59833/diff/1/
>
>
> Testing
> -------
>
> I ran SQL server third party tests with MSSQL 2014 and MSSQL 2017 too.
>
>
> Thanks,
>
> Szabolcs Vasas
>
>