my_workbook

As mentioned in my previous comment:

I'm not sure if this is possible or not. I have spent many hours trying to accomplish the same task. What I now do is manage the data in a separate file and create a connection to that file with the excel file that has the reports and images. In some cases I have written excel macros (called by python) to refresh the data. I nearly always create the connection in power pivot as this makes adding new data much easier. Try out this link here to learn how to connect a flat file to excel's PowerPivot.

I've run your code (I've added some missing fields declaration which you did not provide). But could not reproduce your problem. This is the output excel file.

In my case I'm using maven with the following dependencies:

Use FillPatternType.

Try with this:

Promoting a comment to an answer

This is covered in this Apache POI FAQ Entry - I'm using the poi-ooxml-schemas jar, but my code is failing with "java.lang.NoClassDefFoundError: org/openxmlformats/schemas/something". The key section is:

There are two jar files available, as described in the components overview section. The full jar of all of the schemas is ooxml-schemas-1.3.jar, and it is currently around 15mb. The smaller poi-ooxml-schemas jar is only about 4mb. This latter jar file only contains the typically used parts though.

Many users choose to use the smaller poi-ooxml-schemas jar to save space. However, the poi-ooxml-schemas jar only contains the XSDs and classes that are typically used, as identified by the unit tests. Every so often, you may try to use part of the file format which isn't included in the minimal poi-ooxml-schemas jar. In this case, you should switch to the full ooxml-schemas-1.3.jar. Longer term, you may also wish to submit a new unit test which uses the extra parts of the XSDs, so that a future poi-ooxml-schemas jar will include them.

So, short term you need to swap out your small poi-ooxml-schemas jar for the full ooxml-schemas-1.3 jar. Longer term, if you submit a unit test to Apache POI which uses these extra classes, it'll be included in a future small schema jar.

Maven artifact details for the full schema are covered here on the Apache POI site

The link you provide points to a .xlsx file (but has a .csv extension instead of .xlsx), which is actually a zip file.

That is why you see those strange characters, the first 2 being 'PK', which means it is a zip file.

So you will have to export to .csv first, before using the file.