How to Implement Data Version Management with MDSplus.Tree.setVersionDate for Nodes Using makeSegment to Insert Data

[CuiMo6]

Affiliation
ENN EXL50U

Question
I am currently using MDSplus to implement data version management. At present, I insert data once with the node.putData(np.array([1,2,3])) function, then insert data again with node.putData(np.array([4,5,6])), and I can view data of different versions using MDSplus.Tree.setVersionDate("09-JUN-2026 16:46:00"). However, I have encountered a problem: when I insert data using the node.makeSegment function, close the tree, and then insert data again with node.makeSegment, the new data is appended after the previous data, resulting in two segments of data being inserted. In this case, I cannot use MDSplus.Tree.setVersionDate to view data of different versions. How can I implement version management for segment data?

[CuiMo6]

In [1]: from MDSplus import Tree, Array, Range, Float64, Data

In [2]: t=MDSplus.Tree('my_tree',1,'edit')

NameError Traceback (most recent call last)
Cell In[2], line 1
----> 1 t=MDSplus.Tree('my_tree',1,'edit')

NameError: name 'MDSplus' is not defined

In [3]: t=Tree('my_tree',1,'edit')

In [4]: node=t.getNode(r'\test')

TreeNNF Traceback (most recent call last)
Cell In[4], line 1
----> 1 node=t.getNode(r'\test')

File C:\Program Files\MDSplus\python\MDSplus\tree.py:920, in Tree.getNode(self, name)
918 else:
919 nid = _C.c_int32(0)
--> 920 _exc.checkStatus(
921 _TreeShr._TreeFindNode(self.ctx,
922 _ver.tobytes(str(name)),
923 _C.byref(nid)))
924 return TreeNode(int(nid.value), self)
925 return ans

File C:\Program Files\MDSplus\python\MDSplus\mdsExceptions.py:92, in checkStatus(status, ignore, message)
90 print(exception.message)
91 else:
---> 92 raise exception

TreeNNF: %TREE-W-NNF, Node Not Found

In [5]: t.close()

In [6]: t=Tree('my_tree',1,'edit')

In [7]: node=t.getNode(r'\test')

In [8]: data_to_write = [float(i) for i in range(1000)]

In [9]: segData = Array(data_to_write)

In [10]: start = Float64(10.0)
...: end = Float64(11.0)
...: delta = Float64(1e-3)
...: dim = Range(start, end, delta)

In [11]: node.makeSegment(start, end, dim, segData)

In [12]: t.write()

In [13]: t.close()

In [14]: MDSplus.tcl('set tree my_tree/shot=1')

NameError Traceback (most recent call last)
Cell In[14], line 1
----> 1 MDSplus.tcl('set tree my_tree/shot=1')

NameError: name 'MDSplus' is not defined

In [15]: from MDSplus import Tree, Array, Range, Float64, Data,tcl

In [16]: tcl('set tree my_tree/shot=1')

In [17]: tcl(r'dir/full \test')

\MY_TREE::TOP

:TEST tags: \TEST
Status: on,parent is on, usage signal
compress on put
Data inserted: 09-JUN-2026 18:04:10.25 Owner: uid=0
Dtype: DTYPE_FT Class: CLASS_R Length: 8000 bytes

Total of 1 node.

In [18]: t=Tree('my_tree',1,'edit')

In [19]: node=t.getNode(r'\test')

In [20]: data_to_write = [float(i) for i in range(1000,2000)]

In [21]: segData = Array(data_to_write)

In [22]: node.makeSegment(start, end, dim, segData)

In [23]: t.write()

TreeMOVEERROR Traceback (most recent call last)
Cell In[23], line 1
----> 1 t.write()

File C:\Program Files\MDSplus\python\MDSplus\tree.py:1172, in Tree.write(self)
1168 """Write out edited tree.
1169 @rtype: None
1170 """
1171 with self._lock:
-> 1172 _exc.checkStatus(_TreeShr._TreeWriteTree(self.pctx, 0, 0))

File C:\Program Files\MDSplus\python\MDSplus\mdsExceptions.py:92, in checkStatus(status, ignore, message)
90 print(exception.message)
91 else:
---> 92 raise exception

TreeMOVEERROR: %TREE-E-MOVEERROR, Error replacing original treefile with new one.

In [24]: tcl(r'dir/full \test')

\MY_TREE::TOP

:TEST tags: \TEST
Status: on,parent is on, usage signal
compress on put
Data inserted: 09-JUN-2026 18:07:47.43 Owner: uid=0
Dtype: DTYPE_FT Class: CLASS_R Length: 16000 bytes

Total of 1 node.

In [25]: exit()

C:\Users\yqy20>ipython
Python 3.11.7 (tags/v3.11.7:fa7a6f2, Dec 4 2023, 19:24:49) [MSC v.1937 64 bit (AMD64)]
Type 'copyright', 'credits' or 'license' for more information
IPython 8.19.0 -- An enhanced Interactive Python. Type '?' for help.

In [1]: from MDSplus import Tree, Array, Range, Float64, Data,tcl

In [2]: t=Tree('my_tree',1,'edit')

In [3]: node=t.getNode(r'\test')

In [4]: data_to_write = [float(i) for i in range(1000,2000)]

In [5]: segData = Array(data_to_write)

In [6]: start = Float64(10.0)
...: end = Float64(11.0)
...: delta = Float64(1e-3)
...: dim = Range(start, end, delta)

In [7]: node.makeSegment(start, end, dim, segData)

In [8]: t.write()

In [9]: t.quit()

Additionally, after I edit shot=1, insert data and call t.write(), everything works fine. However, if I call t.close() and then edit again, insert data and call t.write() once more, an error occurs. What is the reason for this?

[mwinkel-dev]

Hi @CuiMo6,

Regarding your two questions . . .

VERSIONING: The makeSegment() feature was not part of the original design of MDSplus; it is a feature that was added later. I hope that nodes with segmented data do support versioning, but won't be surprised if that is not the case. I will do experiments later today and post the results to this issue. If versioning is not supported, I will also suggest some workarounds.

WRITE ERROR: What you describe is unexpected behavior. I will see if I can reproduce the problem. If so, I will submit it as a bug and work on a fix. To investigate this issue, I will configure my development system so it has the same software versions that are in use at your site. Please provide the following information:

• Client computer: versions of Windows, Python and MDSplus used
• Server computer: version of the operating system, Python and MDSplus

[mwinkel-dev]

Hi @CuiMo6 -- I've been experimenting with segmented signals and am fairly confident that they do not support data versioning. Furthermore, a scan of the source code that implements segments didn't find statements that reference data versions.

Note that segments were added to support long running experiments, thus each segment records data for the next period of time. For example, suppose MDSplus is used to record 1 Hz data of room temperature over a 24-hour period with a new segment written every minute. In that scenario, it doesn't make sense to state that the arrival of a new segment indicates a "new" data version.

A different scenario applies to simulation runs. For example, perhaps 4 simulations of plasma current are done before an actual shot is run. Furthermore, suppose that each simulation generates 10 segments of the plasma current signal, and that the actual shot also generates 10 segments. That means there are 5 sets of 10 segments to be stored. MDSplus cannot store all of those segments in a single node (as 5 independent chains of segments). The workaround in that case would either be to create 5 trees (one for each simulation run and one for the shot). Or to have a tree with many nodes (or subtrees) -- one for each simulation.

Regarding the write error you encountered, I am still working on reproducing the issue. It does not occur during the normal workflows I tried that only involve scalar data. However, it is likely that I will be able to reproduce it when I do more experiments with segmented signals.

I will continue to investigate this issue and will post additional findings next week.