Inability to write Dataset to file after decoding

[suever]

From Suever@gmail.com on November 24, 2012 11:28:05

If the user decodes the data elements in a dataset, string and Person Name elements are converted to unicode. When the user attempts to write this dataset to file, unicode encoding isn't handled explicitly so python attempts to implicitly convert to ascii (or another encoding depending upon the user's environment settings). The following code demonstrates this issue:

import dicom from dicom.dataelem import DataElement from dicom.charset import decode from dicom.filewriter import write_string

data_element = DataElement((0x08,0x70),'SH','Suéver') decode(data_element, None) # Decodes using default_encoding write_string(open('/dev/null','wb'), data_element)

NOTE: The é character is not a valid character in ascii therefore write_string will throw an ambiguous UnicodeEncodeError.

This handling of unicode needs to not only be present in filewriter.write_string but also filewriter.write_PN. The handling of PersonName elements should be different due to the fact that PN values use multiple encodings.

I have already begun work on a patch for this, but just wanted to document the issue here in case it crops up in the future.

Original issue: http://code.google.com/p/pydicom/issues/detail?id=118

[suever]

From Suever@gmail.com on November 24, 2012 08:40:03

Status: Started
Labels: Difficulty-Easy

[suever]

From Suever@gmail.com on November 24, 2012 16:30:33

Status: Fixed