9.52 OPG_APIS.SPARSIFY_GRAPH

Format

OPG_APIS.SPARSIFY_GRAPH(
     edge_tab_name IN VARCHAR2,
     threshold     IN NUMBER DEFAULT 0.5,
     min_keep      IN INTEGER DEFAULT 1,
     dop           IN INTEGER DEFAULT 4,
     wt_out_tab    IN OUT VARCHAR2,
     wt_und_tab    IN OUT VARCHAR2,
     wt_hsh_tab    IN OUT VARCHAR2,
     wt_mch_tab    IN OUT VARCHAR2,
     tbs           IN VARCHAR2 DEFAULT NULL,
     options       IN VARCHAR2 DEFAULT NULL);

Description

Performs sparsification (edge trimming) for a property graph edge table.

Parameters

edge_tab_name

Name of the property graph edge table (GE$).

threshold

A numeric value controlling how much sparsification needs to be performed. The lower the value, the more edges will be removed. Some typical values are: 0.1, 0.2, ..., 0.5

min_keep

A positive integer indicating at least how many adjacent edges should be kept for each vertex. A recommended value is 1.

dop

Degree of parallelism for the operation.

wt_out_tab

A working table to hold the output, a sparsified graph.

wt_und_tab

A working table to hold the undirected version of the original graph.

wt_hsh_tab

A working table to hold the min hash values of the graph.

wt_mch_tab

A working table to hold matching count of min hash values.

tbs

A working table to hold the working table data.

options

Additional settings for operation. An optional string with one or more (comma-separated) of the following values:

• 'INMEMORY=T' is an option for creating the schema tables with an 'inmemory' clause.

• 'IMC_MC_B=T' creates the schema tables with an INMEMORY MEMCOMPRESS BASIC clause.

Usage Notes

The CREATE TABLE privilege is required to call this procedure.

The sparsification algorithm used is a min hash based local sparsification. See "Local graph sparsification for scalable clustering", Proceedings of the 2011 ACM SIGMOD International Conference on Management of Data: https://cs.uwaterloo.ca/~tozsu/courses/CS848/W15/presentations/ElbagouryPresentation-2.pdf

Sparsification only involves the topology of a graph. None of the properties (K/V) are relevant.

Examples

The following example does the preparation work for the edges table of mypg, prints out the working table names, and runs sparsification. The output, a sparsified graph, is stored in a table named LEAN_PG, which has two columns, SVID and DVID.

SQL> set serveroutput on
DECLARE
  my_lean_pg  varchar2(100) := 'lean_pg'; -- output table
  wt2 varchar2(100);
  wt3 varchar2(100);
  wt4 varchar2(100);
BEGIN
  opg_apis.sparsify_graph_prep('mypgGE$', my_lean_pg, wt2, wt3, wt4, null);
  dbms_output.put_line('wt2 ' || wt2 || ', wt3 ' || wt3 || ', wt4 '|| wt4);

  opg_apis.sparsify_graph('mypgGE$', 0.5, 1, 4, my_lean_pg, wt2, wt3, wt4, 'SEMTS', null);
END;
/ 
  
wt2 "MYPGGE$$TWSPAU275", wt3 "MYPGGE$$TWSPAH275", wt4 "MYPGGE$$TWSPAM275"


SQL> describe lean_pg;
 Name					   Null?    Type
 ----------------------------------------- -------- ----------------------------
 SVID						    NUMBER
 DVID						    NUMBER