[teiid-issues] [JBoss JIRA] (TEIID-3219) Excel Translator Wrong Result From Columns With Interleaved Null Values

[Ramesh Reddy (JIRA)]

    [ https://issues.jboss.org/browse/TEIID-3219?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13021353#comment-13021353 ] 

Ramesh Reddy commented on TEIID-3219:
-------------------------------------

It will allow all empty lines to be treated as valid lines with all null column values, that means for example a excel sheet can have 10 valid rows values and can have any number of rows that are nothing but nulls after that. The translator can keep looping forever as if it is reading valid values. When I tested, I easily went into millions of iterations with out a valid row. The previous code treated the very first line with a all column rows as null as the end of the sheet. That worked, but the current scenario is there are multiple columns in excel sheet with arbitrary column values as nulls, the table defined on the sheet is only extracting the single column back. Whenever it is seeing the null on the column value it is treating it as the end of file. 

Kylin removed whole end of file concept, which is problematic. I suggested, look ahead on the row cursor for follow on column values to check if they are null or not to define a value is null vs it is end of file. I asked him configure the look-ahead number to be configurable.

> Excel Translator Wrong Result From Columns With Interleaved Null Values
> -----------------------------------------------------------------------
>
>                 Key: TEIID-3219
>                 URL: https://issues.jboss.org/browse/TEIID-3219
>             Project: Teiid
>          Issue Type: Bug
>          Components: Misc. Connectors
>    Affects Versions: 8.7.1
>            Reporter: Jan Stastny
>            Assignee: Kylin Soong
>             Fix For: 8.7.1, 8.10
>
>         Attachments: artifacts.zip
>
>
> When querying a column which has a null value somewhere in it, the result set is missing all the rows under the null value in the xlsx file. When querying more columns, the result set is missing some values only if there is a row, where both columns contains null value.



--
This message was sent by Atlassian JIRA
(v6.3.8#6338)