Pydicom: LookupError: unknown encoding: Not Supplied

[amandagb]

Description

Output from ds = pydicom.read_file(dcmFile) (an RTSTRUCT dicom file, SOP UID 1.2.840.10008.5.1.4.1.1.481.3) results in some tags throwing a LookupError: "LookupError: unknown encoding: Not Supplied"
Specific tags which cannot be decoded are as follows:
['DeviceSerialNumber',
'Manufacturer',
'ManufacturerModelName',
'PatientID',
'PatientName',
'RTROIObservationsSequence',
'ReferringPhysicianName',
'SeriesDescription',
'SoftwareVersions',
'StructureSetLabel',
'StructureSetName',
'StructureSetROISequence',
'StudyDescription',
'StudyID']

I suspect that it's due to the fact that ds.SpecificCharacterSet = 'Not Supplied', but when I try to set ds.SpecificCharacterSet to something reasonable (ie ISO_IR_100 or 'iso8859'), it doesn't seem to make any difference.

Reading the same file, with NO modifications, in gdcm does not result in any errors and all fields are readable.

Steps/Code to Reproduce

import pydicom 
ds = pydicom.read_file(dcmFile)
print(ds.PatientName)

Expected Results

No error is thrown and the name of the patient is printed.

Actual Results

Traceback (most recent call last):
File "", line 1, in
File "C:\Users\Amanda\AppData\Local\Continuum\anaconda3\envs\itk\lib\site-packages\pydicom\valuerep.py", line 706, in __str__
return '='.join(self.components).__str__()
File "C:\Users\Amanda\AppData\Local\Continuum\anaconda3\envs\itk\lib\site-packages\pydicom\valuerep.py", line 641, in components
self._components = _decode_personname(groups, self.encodings)
File "C:\Users\Amanda\AppData\Local\Continuum\anaconda3\envs\itk\lib\site-packages\pydicom\valuerep.py", line 564, in _decode_personname
for comp in components]
File "C:\Users\Amanda\AppData\Local\Continuum\anaconda3\envs\itk\lib\site-packages\pydicom\valuerep.py", line 564, in
for comp in components]
File "C:\Users\Amanda\AppData\Local\Continuum\anaconda3\envs\itk\lib\site-packages\pydicom\charset.py", line 129, in decode_string
return value.decode(encodings[0])
LookupError: unknown encoding: Not Supplied

Versions

Platform: Windows-10-10.0.17763-SP0
Python Version: Python 3.6.4 |Anaconda, Inc.| (default, Mar 12 2018, 20:20:50) [MSC v.1900 64 bit (AMD64)]
pydicom Version: pydicom 1.2.2

[scaramallion]

You said on the pynetdicom issue you can't upload an anonymised file, but can you open the file in a hex editor and post the raw byte output from the first few (non-identifying) elements? From the start of the file to the end of say (0008,0070) should be enough.

Alternatively you could truncate the file at the end of the (0008,0070) element and upload that.

If you need to know how to interpret the encoded data check out Part 5, Chapter 7 of the DICOM Standard. And if the file's been saved in the DICOM File Format there may also be a 128 byte header following by 'DICM' before the start of the dataset (which we don't need).

[darcymason]

open the file in a hex editor and post the raw byte output from the first few (non-identifying) elements

Alternatively:

import pydicom.config
pydicom.config.debug(True)
ds = dcmread(youfilename)

And as suggested copy the first non-identifying part of the debug output for posting.

[amandagb]

Please find the truncated dataset attached as requested. I wasn't allowed to upload a *.dcm file so just wrote it as a *.txt file. The file is readable by pydicom but exhibits the same aforementioned problems, where the "LookupError: unknown encoding: Not Supplied" happens only for the Manufacturer tag.

truncated.txt

[darcymason]

Okay, so I've tried deleting SpecificCharacterSet and the error still occurs. I think pydicom is still holding on to the original values, and needs some code to handle the case when SpecificCharacterSet is deleted or set again after reading the file. We can dig into it a little further.

[scaramallion]

Workaround:

from pydicom import dcmread

ds = dcmread('path/to/file')
del ds.SpecificCharacterSet
ds.read_encoding = []

[amandagb]

Brilliant! That worked!

Thank you for the quick fix! I've found pydicom is a lot more user-friendly than gdcm so I'm really glad I don't have to resort to that. 😀

[scaramallion]

No problem, @mrbean-bremen do you want to handle the underlying issue?

[mrbean-bremen]

Certainly - as I may have introduced it... Not sure if I'll find the time today though.