df034_SaveGraph.py File Reference

Detailed Description

Basic SaveGraph usage.

This tutorial shows how to use the SaveGraph action. SaveGraph inspects the sequence of RDataFrame actions.

 
# First, an RDataFrame computation graph is created with Defines, Filters and methods such as Mean, Count, etc.
# After that, SaveGraph can be called either on the root RDataFrame object or on a specific node of the computation
# graph: in the first case, the graph returned will span the full computation graph, in the second case it will show
# only the branch of the computation graph that the node belongs to.
# If a filename is passed as second argument, the graph is saved to that file, otherwise it is returned as a string.
 
import ROOT
 
rd1 = ROOT.RDataFrame(2)
 
rd2 = rd1.Define("Root_def1", "1") \
         .Filter("Root_def1 < 2", "Main_Filter") \
         .Define("Root_def2", "1")
 
branch1 = rd2.Define("Branch_1_def", "1")
branch2 = rd2.Define("Branch_2_def", "1")
 
branch1_1 = branch1.Filter("Branch_1_def < 2", "Filter_1") \
                   .Define("Branch_1_1_def", "1") \
                   .Filter("1 == Branch_1_1_def % 2", "Filter_1_1") \
                   .Mean("Branch_1_1_def");
 
branch1_2 = branch1.Define("Branch_1_2_def", "1") \
                   .Filter("Branch_1_2_def < 2", "Filter_1_2") \
                   .Count()
 
branch2_1 = branch2.Filter("Branch_2_def < 2", "Filter_2") \
                   .Define("Branch_2_1_def", "1") \
                   .Define("Branch_2_2_def", "1") \
                   .Filter("1 == Branch_2_1_def % 2", "Filter_2_1") \
                   .Max("Branch_2_1_def")
 
branch2_2 = branch2.Count()
 
print(ROOT.RDF.SaveGraph(branch1_1))
ROOT.RDF.SaveGraph(rd1, "exampleGraph.dot")

digraph {
   8 [label="Mean", style="filled", fillcolor="#e47c7e", shape="box"];
   6 [label="Filter_1_1", style="filled", fillcolor="#0f9d58", shape="hexagon"];
   7 [label="Define\nBranch_1_1_def", style="filled", fillcolor="#4285f4", shape="ellipse"];
   3 [label="Filter_1", style="filled", fillcolor="#0f9d58", shape="hexagon"];
   4 [label="Define\nBranch_1_def", style="filled", fillcolor="#4285f4", shape="ellipse"];
   5 [label="Define\nRoot_def2", style="filled", fillcolor="#4285f4", shape="ellipse"];
   1 [label="Main_Filter", style="filled", fillcolor="#0f9d58", shape="hexagon"];
   2 [label="Define\nRoot_def1", style="filled", fillcolor="#4285f4", shape="ellipse"];
   0 [label="Empty source\nEntries: 2", style="filled", fillcolor="#f4b400", shape="ellipse"];
   6 -> 8;
   7 -> 6;
   3 -> 7;
   4 -> 3;
   5 -> 4;
   1 -> 5;
   2 -> 1;
   0 -> 2;
}

Date

January 2022

Author

Ivan Kabadzhov (CERN)

Definition in file df034_SaveGraph.py.