Release History
Help

YMatrix

Quick Start

Deployment

Data Usage

Manage Clusters

Security

Best Practice

Technical Principles

Advanced Features

Graph Database

Guide

SQL Reference

FAQ

Importing Graph Data from Files

You can create a graph from files using the instructions below. This document covers:

  • Functions for loading graphs from files
  • How to enable those functions
  • CSV file structure expected by the loading functions
  • A simple source-code example for loading countries and cities

Users load graphs in two steps:

  • First, load vertices
  • Second, load edges

You must create the graph and its labels before loading data from files.

Loading Functions

The following functions load vertices and edges from files.

The function load_labels_from_file loads vertices from a CSV file.

load_labels_from_file('<graph name>',
                      '<label name>',
                      '<file path>')

A fourth argument lets users omit the id field. Use this option only when the input file contains no id column.

load_labels_from_file('<graph name>',
                      '<label name>',
                      '<file path>',
                      false)

The function load_edges_from_file loads edges from a CSV file. See the file structure specification below.

Note: Ensure that id values in the edge file match existing id values in the vertex file.

load_edges_from_file('<graph name>',
                    '<label name>',
                    '<file path>');

CSV File Format Specification

The following describes the required structure for vertex and edge CSV files.

  • A vertex CSV file must have the following format:
Column Name Description
id Must be the first column. All values must be positive integers. Optional only when id_field_exists is set to false. Otherwise, it must be present.
Properties All remaining columns define vertex properties. The header row must contain property names.
  • An edge CSV file must have the following format:
Column Name Description
start_id The id of the source vertex. Must exist in the vertex file.
start_vertex_type The label of the source vertex.
end_id The id of the target vertex.
end_vertex_type The label of the target vertex.
properties Edge properties. Header row must contain property names.

Example files are located in regress/age_load/data.

SQL Script Examples

  • Load the AGE extension and create a graph:
LOAD 'age';

SET search_path TO ag_catalog;
SELECT create_graph('agload_test_graph');
  • Create the Country label and load vertices from a CSV file. Note: This CSV file includes an id column.
SELECT create_vlabel('agload_test_graph','Country');
SELECT load_labels_from_file('agload_test_graph',
                             'Country',
                             'age/regress/age_load/data/countries.csv');
  • Create the City label and load vertices from a CSV file. Note: This CSV file includes an id column.
SELECT create_vlabel('agload_test_graph','City');
SELECT load_labels_from_file('agload_test_graph',
                             'City',
                             'age/regress/age_load/data/cities.csv');
  • Create the has_city edge label and load edges from a CSV file.
SELECT create_elabel('agload_test_graph','has_city');
SELECT load_edges_from_file('agload_test_graph', 'has_city',
     'age/regress/age_load/data/edges.csv');
  • Verify correct graph loading:
SELECT table_catalog, table_schema, table_name, table_type
FROM information_schema.tables
WHERE table_schema = 'agload_test_graph';

SELECT COUNT(*) FROM agload_test_graph."Country";
SELECT COUNT(*) FROM agload_test_graph."City";
SELECT COUNT(*) FROM agload_test_graph."has_city";

SELECT COUNT(*) FROM cypher('agload_test_graph', $$MATCH(n) RETURN n$$) AS (n agtype);
SELECT COUNT(*) FROM cypher('agload_test_graph', $$MATCH (a)-[e]->(b) RETURN e$$) AS (n agtype);

Creating Vertices When the Input File Has No id Column

  • Create the Country2 label and load vertices from a CSV file. Note: This CSV file has no id column.
SELECT create_vlabel('agload_test_graph','Country2');
SELECT load_labels_from_file('agload_test_graph',
                             'Country2',
                             'age/regress/age_load/data/countries.csv',
                             false);
  • Create the City2 label and load vertices from a CSV file. Note: This CSV file has no id column.
SELECT create_vlabel('agload_test_graph','City2');
SELECT load_labels_from_file('agload_test_graph',
                             'City2',
                             'age/regress/age_load/data/cities.csv',
                             false);
  • Verify correct graph loading and compare auto-generated IDs versus IDs read from the file:
    • Labels Country and City use id values from the input file.
    • Labels Country2 and City2 were created without an id column — IDs were auto-assigned.
SELECT COUNT(*) FROM agload_test_graph."Country2";
SELECT COUNT(*) FROM agload_test_graph."City2";

SELECT id FROM agload_test_graph."Country" LIMIT 10;
SELECT id FROM agload_test_graph."Country2" LIMIT 10;

SELECT * FROM cypher('agload_test_graph', $$MATCH(n:Country {iso2 : 'BE'})
    RETURN id(n), n.name, n.iso2 $$) AS ("id(n)" agtype, "n.name" agtype, "n.iso2" agtype);
SELECT * FROM cypher('agload_test_graph', $$MATCH(n:Country2 {iso2 : 'BE'})
    RETURN id(n), n.name, n.iso2 $$) AS ("id(n)" agtype, "n.name" agtype, "n.iso2" agtype);

SELECT * FROM cypher('agload_test_graph', $$MATCH(n:Country {iso2 : 'AT'})
    RETURN id(n), n.name, n.iso2 $$) AS ("id(n)" agtype, "n.name" agtype, "n.iso2" agtype);
SELECT * FROM cypher('agload_test_graph', $$MATCH(n:Country2 {iso2 : 'AT'})
    RETURN id(n), n.name, n.iso2 $$) AS ("id(n)" agtype, "n.name" agtype, "n.iso2" agtype);

SELECT drop_graph('agload_test_graph', true);
← Previous
Comparability, Equality, Orderability and Equivalence
Next →
CREATE