Error : a bytes-like object is required, not 'MultiValue'

[nizarbabu]

Hello,

I am getting following error while updating the tag LongTrianglePointIndexList (0066,0040),
TypeError: a bytes-like object is required, not 'MultiValue'

I noticed that the error gets produced only when the VR is given as "OL" , works fine with "OB", "OF" etc.

sample code (assume 'lineSeq' is the dicom dataset sequence):

import pydicom
import array
data=list(range(1,10))
data=array.array('H', indexData).tostring()  # to convert to unsigned short
lineSeq.add_new(0x00660040, 'OL', data)   
ds.save_as("mydicom")

outcome: TypeError: a bytes-like object is required, not 'MultiValue'

using version - 2.0.0.0

Any help is appreciated.

Thank you

[nizarbabu]

Also tried following code to get the byte string, but same error.

1. data=array.array('L', indexData).tostring() # to convert to long -> same error
2. data=array.array('Q', indexData).tostring() # to convert to long long -> same error

[scaramallion]

O* VRs should be bytes. Use array.tobytes() instead of tostring()?

Also, in the future if have an issue it's much more helpful if you post the full traceback rather than the error since we can look at it to figure out where in the code the exception is occurring.

It would also help if you posted the version of Python you're using.

This works fine for me with Python 3.9 and pydicom 2.1.2:

from pydicom import Dataset
import array

arr = array.array('H', range(10))
ds = Dataset()
ds.is_little_endian = True
ds.is_implicit_VR = False
ds.LongTrianglePointIndexList = arr.tobytes()
print(ds["LongTrianglePointIndexList"].VR)  # 'OL'
ds.save_as('temp.dcm')

This also works fine:

ds = Dataset()
ds.add_new(0x00660040, 'OL', arr.tobytes())

[nizarbabu]

Thank you for the answer.
Unfortunately the error still persists with above code.
Please find the attached detailed error.
error.txt

One more information is that the 'ds' is actually read from a file in the disk (ds=pydicom.read_file(filename)).
and this byte array is stored under the following sequence
ds[0x0066,0x0002][0][0x0066,0x0013][0][0x0066,0x0028][0][0x0066,0x0040] = arr.tobytes()

pydicom - 2.0.0.0
python - 3.6.4

Thank you.

[scaramallion]

Could you post a minimal code sample that reproduces the issue please?

If you're using something like this:
ds[0x0066,0x0002][0][0x0066,0x0013][0][0x0066,0x0028][0][0x0066,0x0040] = arr.tobytes()

Then you're missing the .value assignment:
ds[0x0066,0x0002][0][0x0066,0x0013][0][0x0066,0x0028][0][0x0066,0x0040].value = arr.tobytes()

[nizarbabu]

Hello,
above code line I just mentioned to give an idea where the actual data is stored (tree level).

Please find the actual code used below,

import pydicom
from pydicom.sequence import Sequence
from pydicom.dataelem import DataElement
from pydicom.dataset import Dataset

ds = pydicom.read_file(filename)
surfaceSeq= ds[0x0066,0x0002]

#// read existing sequence items in the dataset
seqlist=[]
for n in surfaceSeq:
    seqlist.append(n)

newDs = Dataset()
 
surfaceMeshPrimitiveSq = Dataset()
lineSeq = Dataset()
indexData = list(range(1,100))
indexData = array.array('H', indexData)
indexData = indexData.tobytes()
lineSeq.add_new(0x00660040, 'OL', indexData) 
surfaceMeshPrimitiveSq.add_new(0x00660028, 'SQ', [lineSeq])
newDs.add_new(0x00660013, 'SQ', [surfaceMeshPrimitiveSq])

#add the new sequnce item to the list
seqlist.append(newDs)
ds[0x0066,0x0002] = DataElement(0x00660002,"SQ",seqlist)
ds.save_as(filename)

[scaramallion]

OK, I can reproduce with:

import array

from pydicom import Dataset
from pydicom.uid import ExplicitVRLittleEndian

ds = Dataset()
ds.file_meta = Dataset()
ds.file_meta.TransferSyntaxUID = ExplicitVRLittleEndian

b = array.array('H', range(100)).tobytes()

ds.LongPrimitivePointIndexList = b
ds.save_as('1421.dcm')

And print(ds) gives:

(0066, 0040) Long Primitive Point Index List     OL: [b'\x00\x00\x01\x00\x02\x00\x03\x00\x04\x00\x05\x00\x06\x00\x07\x00\x08\x00\t\x00\n\x00\x0b\x00\x0c\x00\r\x00\x0e\x00\x0f\x00\x10\x00\x11\x00\x12\x00\x13\x00\x14\x00\x15\x00\x16\x00\x17\x00\x18\x00\x19\x00\x1a\x00\x1b\x00\x1c\x00\x1d\x00\x1e\x00\x1f\x00 \x00!\x00"\x00#\x00$\x00%\x00&\x00\'\x00(\x00)\x00*\x00+\x00,\x00-\x00.\x00/\x000\x001\x002\x003\x004\x005\x006\x007\x008\x009\x00:\x00;\x00<\x00=\x00>\x00?\x00@\x00A\x00B\x00C\x00D\x00E\x00F\x00G\x00H\x00I\x00J\x00K\x00L\x00M\x00N\x00O\x00P\x00Q\x00R\x00S\x00T\x00U\x00V\x00W\x00X\x00Y\x00Z\x00[\x00', b'\x00]\x00^\x00_\x00`\x00a\x00b\x00c\x00']

I think this is because the byte value is hitting the hex for the backslash character during assignment. Ouch, that's kinda nasty.

[scaramallion]

As a workaround you can do the equivalent of:

b = array.array('H', range(100)).tobytes()
ds.LongPrimitivePointIndexList = None
# Bypass the value check and set it directly
ds["LongPrimitivePointIndexList"]._value = b

The problem arises because elements with a VR of OL (and OD and OV) aren't being excluded from the backslash check. Thanks, we'll have a fix in for the next release.

[nizarbabu]

Thank you.
The workaround works!

[scaramallion]

We really should have a checklist or something for what needs to be changed when we add new VRs...