Working With Trees
The trees which contain the data pertaining to an experiment or simulation are stored on an MDSplus server as files. Each tree consists of a model (shot # -1) and one or more pulse files with positive shot numbers. Special shot numbers:
- -1 - model
- 0 - current shot
- >1 - pulse files
The MDSplus data system is based around hierarchical data stores called trees. A tree is a collection of nodes which describe something. All data in MDSplus are stored in trees. Models vs Shots Trees vs Subtrees
Opening vs Editing
Files and treenames
What are the files called ? Where are they stored ? What is in them ? MDSplus trees are stored on disk in a set of three files. Their base names are name-of-the-tree_shot-specifier where the shot-specifier is either a three or 10 digit shot number, or the string model. For example the files for a tree called demo would be called: demo_001.* for shot 1, demo_1010000001.* for shot 1010000001 and demo_model for the model (shot -1).
|treename_nnn.tree||This file contains the structure of the tree. This file is opened readonly unless the tree is being edited. It contains the names of the nodes, thier locations in the hierarchy, and their usages.|
|treename_nnn.characteristics||This file contains a set of fixed length properties for each node. This file is opened for shared read and write access. Including the datatype, data shape, when it was last written, who last wrote it, and the location of the data record in the datafile.|
|treename_nnn.datafile||This file contains the actual data. This file is opend for shared read and write access.|
Environment variables define the location of the files which make up a given The files for a each tree are located on disk by the translation of a name treename_path. Organizing data Tags What they are, How to create them (tcl commands and traverser), environment variables/registry, model/pulse file concepts.... You should also probably discuss local/thick/distributed trees and file name masks.