[PDFBOX-4749] Text Extraction leads to weird result - toUnicodeCMap is 'AdHoc-UCS'

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 2.0.18
• Fix Version/s: 3.0.0 PDFBox
• Component/s: Text extraction
• Assignee: Andreas Lehmkühler
• Reporter: Benoit Lacelle
• Created: 21/Jan/20 09:50
• Updated: 18/Aug/23 05:46
• Resolved: 03/Nov/20 16:24

Description

I consider the attached PDF. I consider the text on the first page:

"Am Fährweg"

It appears the code for the first character 'A' is 65 and is parsed correctly, while the code for the fourth character 'F' is 70 which is parsed as a 'c'.

org.apache.pdfbox.pdmodel.font.PDFont.toUnicode(int) relies on a CMap named 'AdHoc-UCS' which mapping in :

{129=ü, 3= , 8=%, 9=&, 11=(, 12=), 15=,, 16=-, 17=., 18=/, 19=0, 20=1, 21=2, 22=3, 23=4, 24=5, 25=6, 26=7, 27=8, 28=9, 29=:, 34=?, 36=A, 37=B, 38=C, 39=D, 40=E, 41=F, 42=G, 43=H, 44=I, 46=K, 47=L, 48=M, 49=N, 50=O, 51=P, 53=R, 54=S, 55=T, 56=U, 57=V, 58=W, 59=X, 61=Z, 68=a, 69=b, 70=c, 71=d, 72=e, 73=f, 74=g, 75=h, 76=i, 78=k, 79=l, 80=m, 81=n, 82=o, 83=p, 85=r, 86=s, 87=t, 88=u, 89=v, 90=w, 93=z, 95=|, 108=ä, 124=ö}

-> 'A' is parsed as 'A' as it is out of the mapping of CMap, while 'F' conflicts the entry mapping 70 to c.

The document is correctly parsed in Acrobat Reader.

Issue Links

• links to: Stackoverflow - PDFBox 2.0.7 ExtractText not working but 1.8.13 does and PDFReader as well