I hope I am addressing the right list. I am an Apache NIFI developer and I
have developed an Apache NIFI processor named SplitExcel which splits a
multi sheet Microsoft Excel spreadsheet into multiple Microsoft Excel
spreadsheets where each sheet from the original file is converted to an
individual spreadsheet. Source code for this can be found here
<https://github.com/apache/nifi/blob/main/nifi-extension-bundles/nifi-poi-bundle/nifi-poi-services/src/main/java/org/apache/nifi/processors/excel/SplitExcel.java>
(lines
165-188).
In order to accomplish this, I use the
com.github.pjfanning.xlsx.StreamingReader to stream the Excel workbook
which allows for easy looping over each sheet. I create a new XSSFWorkbook
for each sheet there is in the existing workbook and I use XSSFSheet
copyRows method to copy the rows from the existing sheet to a new sheet to
place in the new XSSFWorkbook. The CellCopyPolicy used has cellStyle set to
false in order to avoid exceeding the maximum number of cell styles (fixed
in NIFI-13726 <https://issues.apache.org/jira/browse/NIFI-13726>). The
drawback though of using cellStyle set to false is that any columns which
have dates, its date formatting is lost and what is left is a meaningless
number (reported in NIFI-14106
<https://issues.apache.org/jira/browse/NIFI-14106>). Is there a way I can
use copyRows not to exceed the maximum number of cell styles and yet have
the formatting needed for dates?
