# Dicomizer - json structure for dataset.json file
**URL:** https://discourse.orthanc-server.org/t/dicomizer-json-structure-for-dataset-json-file/755
**Category:** Google Groups archive
**Created:** [August 17, 2017, 6:31pm UTC](https://discourse.orthanc-server.org/t/dicomizer-json-structure-for-dataset-json-file/755 "2017-08-17T18:31:23Z")
**Posts on this page:** 4
**Page:** 1
### Author:  [@Jack\_Daniels](https://discourse.orthanc-server.org/u/Jack_Daniels)
#### Post date: [August 17, 2017, 6:31pm UTC](https://discourse.orthanc-server.org/t/dicomizer-json-structure-for-dataset-json-file/755/1 "2017-08-17T18:31:23Z")
I am trying to create DICOM images from WSI-tif files for upload into orthanc. Since the sample json-config file does not all items I would like to include in the DICOM file, I tried editing the sample\_dataset.json file. Unfortunately, I cannot figure out the correct format - no matter what I do,
the parsing of the modified json file always fails.
Is there a place/document that explains how to add items to the default file (what are the allowed keywords and sequence structures)?
Thank you very much - Jack
---
### Author:  [@Thibault\_Nelis](https://discourse.orthanc-server.org/u/Thibault_Nelis)
#### Post date: [August 18, 2017, 8:03am UTC](https://discourse.orthanc-server.org/t/dicomizer-json-structure-for-dataset-json-file/755/2 "2017-08-18T08:03:47Z")
Hello,
> Is there a place/document that explains how to add items to the
> default file (what are the allowed keywords and sequence structures)?
Good question, I don't think so actually.
If you can read C++, the parser is here[1] and more specifically
there[2]. If not then you should file a request on the issue tracker
for someone to synthesize its behavior.
You may also send your dataset, we might try to take a look at it
specifically to see what's wrong.
[1] [https://bitbucket.org/sjodogne/orthanc-wsi/src/42dcf14389434478bd8e](https://bitbucket.org/sjodogne/orthanc-wsi/src/42dcf14389434478bd8e)
4471e6fbe8770a966674/Applications/Dicomizer.cpp?at=default&fileviewer=f
ile-view-default#Dicomizer.cpp-203
[2] [https://bitbucket.org/sjodogne/orthanc-wsi/src/42dcf14389434478bd8e](https://bitbucket.org/sjodogne/orthanc-wsi/src/42dcf14389434478bd8e)
4471e6fbe8770a966674/Resources/Orthanc/OrthancServer/FromDcmtkBridge.cp
p?at=default&fileviewer=file-view-default#FromDcmtkBridge.cpp-1657
---
### Author:  [@jodogne](https://discourse.orthanc-server.org/u/jodogne)
#### Post date: [August 18, 2017, 11:17am UTC](https://discourse.orthanc-server.org/t/dicomizer-json-structure-for-dataset-json-file/755/3 "2017-08-18T11:17:17Z")
Dear Jack,
The “dataset.json” format to be used with the “–dataset” command-line argument is really simple: Just add all the tags you want to be included in the output DICOM files into the root associative JSON array. Check out the related documentation:
[http://book.orthanc-server.com/plugins/wsi.html#specifying-a-dicom-dataset](http://book.orthanc-server.com/plugins/wsi.html#specifying-a-dicom-dataset)
A sample JSON file can found at the following location (it is the sample dataset that is generated by adding the “–sample-dataset” flag when invoking “OrthancWSIDicomizer”):
[https://bitbucket.org/sjodogne/orthanc-wsi/src/default/Resources/SampleDataset.json](https://bitbucket.org/sjodogne/orthanc-wsi/src/default/Resources/SampleDataset.json)
The tags can be specified either with their symbolic tag name, or with their hexadecimal identifier. For instance:
{
“PatientID” : “C123456789”,
“0010,0010” : “PATIENT^NAME”
}
Please share a buggy JSON dataset so as further help can be provided.
Regards,
Sébastien-
---
### Author:  [@Jack\_Daniels](https://discourse.orthanc-server.org/u/Jack_Daniels)
#### Post date: [August 18, 2017, 11:37pm UTC](https://discourse.orthanc-server.org/t/dicomizer-json-structure-for-dataset-json-file/755/4 "2017-08-18T23:37:58Z")
Sébastien, Thank you very much for you comments, in particular the possibility of using Byte-codes "0040, 0052", etc. That made things much easier. The original problem was caused by putting an element in the wrong sequence - my mistake.
Thank you very much,
Jack