Hi,
users mailing list is fine. Please share your file by uploading it to a
sharehoster.
Yes you can use the ExtractTextByArea class, or alternatively set
"beads" (kindof invisible rectangles) on top of the pages and then
extract normally, however both approaches require you to know where the
columns are.
Tilman
Am 17.04.2026 um 22:53 schrieb Robert Rodini:
I have used CLI PDFBox utility successfully for years on a third party PDF
which is issued twice a year. The PDF is always produced with a 3 column format
and the extracted text always comes out column by column from top to bottom
from each column. That is until now.
Not the 3rd party changed the internals of the PDF such that PDFBox extracts
the text in a somewhat unpredictable order. It seems to work left to right
horizontally multiple times. The extracted text in no longer in the expected
order.
Can you steer me to the PDFBox APIs that might help me understand the new
internal structure? My initial goal is to write a Java program that can
distinguish the old PDF files from the new PDF files. Later, to write my own
extraction program.
Thank you
P.S. Should this question be submitted [email protected]?
