JDBC data types from org.postgresql.geometric are treated as referenced agreggates instead of jdbc data types

[iarobles]

There seems to be a problem when save or findById methods of the CrudRepository are called on an entity that has a property of the type in org.postgresql.geometric (for example PGpoint). This types are used to map postgres geometric data types: https://www.postgresql.org/docs/9.4/datatype-geometric.html. A similar problem happens with any of the PostGIS JDBC types.

The error seems to be due to Spring Data JDBC treating the property as if it were a referenced aggregate instead of a data type describing a column data type. For example, for an entity with the following properties:

@Table("example")
public class Example {
    @Id
    private int id;
    private String info;
    private PGpoint singlePoint;

    public int getId() {return id;}
    public void setId(int id) {this.id = id;}
    public String getInfo() {return info;}
    public void setInfo(String info) {this.info = info;}
    public PGpoint getSinglePoint() { return singlePoint;}
    public void setSinglePoint(PGpoint singlePoint) {this.singlePoint = singlePoint;}
}

when save is called on an instance of this class, the following error happens (debug level):

INSERT INTO "p_gpoint" ("example", "is_null", "type", "value", "x", "y") VALUES (?, ?, ?, ?, ?, ?)] for task [PreparedStatementCallback

Note that "x", "y", "isNull", etc. are public properties of the PGPoint class: https://github.com/pgjdbc/pgjdbc/blob/master/pgjdbc/src/main/java/org/postgresql/geometric/PGpoint.java

Let me know if you need more information and thanks in advance!

[mp911de]

Any complex objects are considered to represent references. With version 2.3 we introduced JdbcPostgresDialect that allows registration of simple types that are not converted. We have a similar arrangement in R2DBC.