[
https://issues.apache.org/jira/browse/TIKA-3030?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17025586#comment-17025586
]
Nick Burch commented on TIKA-3030:
----------------------------------
Pretty sure we've got a test file in Apache POI like this - some third party
tools generate with the wrong name. I've got a feeling that
{{WORKBOOK_in_capitals.xls}} will be one such file, but I'm on the wrong device
to check right now!
See
[https://github.com/apache/poi/blob/trunk/src/java/org/apache/poi/hssf/model/InternalWorkbook.java#L93]
for the list of well-ish known variants
> XLS files with a root node named WORKBOOK don't get parsed
> ----------------------------------------------------------
>
> Key: TIKA-3030
> URL: https://issues.apache.org/jira/browse/TIKA-3030
> Project: Tika
> Issue Type: Bug
> Components: parser
> Affects Versions: 1.23
> Reporter: Clark Perkins
> Priority: Major
>
> I have an XLS file where the root node contains 2 top-level names -
> "WORKBOOK" and " SummaryInformation".
> The type gets properly detected as "application/vnd.ms-excel", because the
> POIFSContainerDetector does a check like so:
> {noformat}
> if (names.contains("Workbook") || names.contains("WORKBOOK")) {
> ...
> }{noformat}
> However, the ExcelExtractor silently rejects the file because the root node
> doesn't contain a top level node named "Workbook".
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
