I tried calling PdfReader on a pdf, but got the error.

Environment

Which environment were you using when you encountered the problem?

$ python -m platform
Linux-4.14.314-164.539.amzn1.x86_64-x86_64-with-glibc2.17

$ python -c "import pypdf;print(pypdf.__version__)"
3.13.0

Traceback

This is the complete Traceback I see:

entry 5928 in Xref table invalid but object found

TypeError Traceback (most recent call last)
Cell In[36], line 1
----> 1 PdfReader('temp/t.pdf')

File ~/SageMaker/custom-miniconda/miniconda/envs/py311/lib/python3.11/site-packages/pypdf/_reader.py:326, in PdfReader.init(self, stream, strict, password)
324 with open(stream, "rb") as fh:
325 stream = BytesIO(fh.read())
--> 326 self.read(stream)
327 self.stream = stream
329 self._override_encryption = False

File ~/SageMaker/custom-miniconda/miniconda/envs/py311/lib/python3.11/site-packages/pypdf/_reader.py:1548, in PdfReader.read(self, stream)
1545 logger_warning(f"incorrect startxref pointer({xref_issue_nr})", name)
1547 # read all cross reference tables and their trailers
-> 1548 self._read_xref_tables_and_trailers(stream, startxref, xref_issue_nr)
1550 # if not zero-indexed, verify that the table is correct; change it if necessary
1551 if self.xref_index and not self.strict:

File ~/SageMaker/custom-miniconda/miniconda/envs/py311/lib/python3.11/site-packages/pypdf/_reader.py:1758, in PdfReader._read_xref_tables_and_trailers(self, stream, startxref, xref_issue_nr)
1756 x = stream.read(1)
1757 if x == b"x":
-> 1758 startxref = self._read_xref(stream)
1759 elif xref_issue_nr:
1760 try:

File ~/SageMaker/custom-miniconda/miniconda/envs/py311/lib/python3.11/site-packages/pypdf/_reader.py:1794, in PdfReader._read_xref(self, stream)
1793 def _read_xref(self, stream: StreamType) -> Optional[int]:
-> 1794 self._read_standard_xref_table(stream)
1795 read_non_whitespace(stream)
1796 stream.seek(-1, 1)

File ~/SageMaker/custom-miniconda/miniconda/envs/py311/lib/python3.11/site-packages/pypdf/_reader.py:1661, in PdfReader._read_standard_xref_table(self, stream)
1659 stream.seek(-1, 1)
1660 cnt = 0
-> 1661 while cnt < size:
1662 line = stream.read(20)
1664 # It's very clear in section 3.4.3 of the PDF spec
1665 # that all cross-reference table lines are a fixed
1666 # 20 bytes (as of PDF 1.7). However, some files have
1667 # 21-byte entries (or more) due to the use of \r\n
1668 # (CRLF) EOL's. Detect that case, and adjust the line
1669 # until it does not begin with a \r (CR) or \n (LF).

TypeError: '<' not supported between instances of 'int' and 'NameObject'

Note

I can't share the pdf.