Originally written by Peter T. Mount (peter@retep.org.uk), the original author of the JDBC driver. JDBC is a core API of Java 1.1 and later. It provides a standard set of interfaces to SQL-compliant databases. Postgres provides a type 4 JDBC Driver. Type 4 indicates that the driver is written in Pure Java, and communicates in the database system's own network protocol. Because of this, the driver is platform independent; once compiled, the driver can be used on any system. This chapter is not intended as a complete guide to JDBC programming, but should help to get you started. For more information refer to the standard JDBC API documentation. Also, take a look at the examples included with the source. The basic example is used here. Precompiled versions of the driver can be downloaded from the PostgreSQL JDBC web site. Alternatively you can build the driver from source. Although you should only need to do this if you are making changes to the source code. Starting with PostgreSQL version 7.1, the JDBC driver is built using Ant, a special tool for building Java-based packages. You should download Ant from the Ant web site and install it before proceeding. Precompiled Ant distributions are typically set up to read a file .antrc in the current user's home directory for configuration. For example, to use a different JDK than the default, this may work: JAVA_HOME=/usr/local/sun-jdk1.3 JAVACMD=$JAVA_HOME/bin/java To build the driver, add the --with-java option to your configure command line, e.g., $ ./configure --prefix=xxx --with-java ... This will build and install the driver along with the rest of the PostgreSQL package when you issue the make/gmake and make/gmake install commands. If you only want to build the driver and not the rest of PostgreSQL, change into the directory src/interfaces/jdbc and issue the respective make/gmake command there. Refer to the PostgreSQL installation instructions for more information about the configuration and build process. When building the driver from source the jar file that is created will be named postgresql.jar. The build will create this file in the src/interfaces/jdbc/jars directory. The resulting driver will be built for the version of Java you are running. If you build with a 1.1 JDK you will build a version that supports the jdbc1 specification, if you build with a Java2 JDK (i.e. JDK1.2 or JDK1.3) you will build a version that supports the jdbc2 specification. Do not try to build the driver by calling javac directly, as the driver uses some dynamic loading techniques for performance reasons, and javac cannot cope. Do not try to run ant directly either, because some configuration information is communicated through the makefiles. Running ant directly without providing these parameters will result in a broken driver. To use the driver, the jar archive (named postgresql.jar if you built from source, otherwise it will likely be named jdbc7.2-1.1.jar or jdbc7.2-1.2.jar for the jdbc1 and jdbc2 versions respectively) needs to be included in the class path, either by putting it in the CLASSPATH environment variable, or by using flags on the java command line. By default, the jar archive is installed in the directory /usr/local/pgsql/share/java. You may have it in a different directory if you used the --prefix option when you ran configure, or if you are using a binary distribution that places it in some different location. For instance, I have an application that uses the JDBC driver to access a large database containing astronomical objects. I have the application and the JDBC driver installed in the /usr/local/lib directory, and the Java JDK installed in /usr/local/jdk1.3.1. To run the application, I would use: export CLASSPATH=/usr/local/lib/finder.jar:/usr/local/pgsql/share/java/postgresql.jar:. java Finder finder.jar contains the Finder application. Loading the driver from within the application is covered in . Because Java only uses TCP/IP connections, the PostgreSQL server must be configured to accept TCP/IP connections. This can be done by setting tcpip_socket = true in the postgresql.conf file or by supplying the -i option flag when starting postmaster. Also, the client authentication setup in the pg_hba.conf file may need to be configured. Refer to the Administrator's Guide for details. The JDBC Driver supports trust, ident, password, md5, and crypt authentication methods. Any source that uses JDBC needs to import the java.sql package, using: import java.sql.*; Do not import the org.postgresql package. If you do, your source will not compile, as javac will get confused. Before you can connect to a database, you need to load the driver. There are two methods available, and it depends on your code which is the best one to use. In the first method, your code implicitly loads the driver using the Class.forName() method. For Postgres, you would use: Class.forName("org.postgresql.Driver"); This will load the driver, and while loading, the driver will automatically register itself with JDBC. The forName() method can throw a ClassNotFoundException if the driver is not available. This is the most common method to use, but restricts your code to use just Postgres. If your code may access another database system in the future, and you do not use any Postgres-specific extensions, then the second method is advisable. The second method passes the driver as a parameter to the JVM as it starts, using the -D argument. Example: java -Djdbc.drivers=org.postgresql.Driver example.ImageViewer In this example, the JVM will attempt to load the driver as part of its initialization. Once done, the ImageViewer is started. Now, this method is the better one to use because it allows your code to be used with other database packages without recompiling the code. The only thing that would also change is the connection URL, which is covered next. One last thing: When your code then tries to open a Connection, and you get a No driver available SQLException being thrown, this is probably caused by the driver not being in the class path, or the value in the parameter not being correct. With JDBC, a database is represented by a URL (Uniform Resource Locator). With Postgres, this takes one of the following forms: jdbc:postgresql:database jdbc:postgresql://host/database jdbc:postgresql://host:port/database where: host The host name of the server. Defaults to localhost. port The port number the server is listening on. Defaults to the Postgres standard port number (5432). database The database name. To connect, you need to get a Connection instance from JDBC. To do this, you would use the DriverManager.getConnection() method: Connection db = DriverManager.getConnection(url, username, password); To close the database connection, simply call the close() method to the Connection: db.close(); Any time you want to issue SQL statements to the database, you require a Statement instance. Once you have a Statement, you can use the executeQuery() method to issue a query. This will return a ResultSet instance, which contains the entire result. illustrates this process. This example with issue a simple query and print out the first column of each row. Statement st = db.createStatement(); ResultSet rs = st.executeQuery("SELECT * FROM mytable"); while(rs.next()) { System.out.print("Column 1 returned "); System.out.println(rs.getString(1)); } rs.close(); st.close(); The following must be considered when using the Statement interface: You can use a single Statement instance as many times as you want. You could create one as soon as you open the connection and use it for the connection's lifetime. But you have to remember that only one ResultSet can exist per Statement. If you need to perform a query while processing a ResultSet, you can simply create and use another Statement. If you are using threads, and several are using the database, you must use a separate Statement for each thread. Refer to if you are thinking of using threads, as it covers some important points. The following must be considered when using the ResultSet interface: Before reading any values, you must call next(). This returns true if there is a result, but more importantly, it prepares the row for processing. Under the JDBC specification, you should access a field only once. It is safest to stick to this rule, although at the current time, the Postgres driver will allow you to access a field as many times as you want. You must close a ResultSet by calling close() once you have finished using it. Once you make another query with the Statement used to create a ResultSet, the currently open ResultSet instance is closed automatically. To perform an update (or any other SQL statement that does not return a result), you simply use the executeUpdate() method: st.executeUpdate("CREATE TABLE basic (a int, b int)"); In Postgres, Large Objects (also known as BLOBs) are used to hold data in the database that cannot be stored in a normal SQL table. They are stored in a separate table in a special format, and are referred to from your own tables by an OID value. For Postgres, you must access Large Objects within an SQL transaction. You would open a transaction by using the setAutoCommit() method with an input parameter of false: Connection mycon; ... mycon.setAutoCommit(false); ... // now use Large Objects There are two methods of using Large Objects. The first is the standard JDBC way, and is documented here. The other, uses PostgreSQL extensions to the API, which presents the libpq large object API to Java, providing even better access to large objects than the standard. Internally, the driver uses the extension to provide large object support. In JDBC, the standard way to access Large Objects is using the getBinaryStream() method in ResultSet, and setBinaryStream() method in PreparedStatement. These methods make the large object appear as a Java stream, allowing you to use the java.io package, and others, to manipulate the object. illustrates the usage of this approach. For example, suppose you have a table containing the file name of an image and you have a large object containing that image: CREATE TABLE images (imgname text, imgoid oid); To insert an image, you would use: File file = new File("myimage.gif"); FileInputStream fis = new FileInputStream(file); PreparedStatement ps = conn.prepareStatement("INSERT INTO images VALUES (?, ?)"); ps.setString(1, file.getName()); ps.setBinaryStream(2, fis, file.length()); ps.executeUpdate(); ps.close(); fis.close(); The question marks must appear literally. The actual data is substituted by the next lines. Here, setBinaryStream transfers a set number of bytes from a stream into a Large Object, and stores the OID into the field holding a reference to it. Notice that the creation of the Large Object itself in the database happens transparently. Retrieving an image is even easier. (We use PreparedStatement here, but the Statement class can equally be used.) PreparedStatement ps = con.prepareStatement("SELECT imgoid FROM images WHERE imgname=?"); ps.setString(1, "myimage.gif"); ResultSet rs = ps.executeQuery(); if (rs != null) { while(rs.next()) { InputStream is = rs.getBinaryStream(1); // use the stream in some way here is.close(); } rs.close(); } ps.close(); Here you can see how the Large Object is retrieved as an InputStream. You will also notice that we close the stream before processing the next row in the result. This is part of the JDBC specification, which states that any InputStream returned is closed when ResultSet.next() or ResultSet.close() is called. Postgres is an extensible database system. You can add your own functions to the backend, which can then be called from queries, or even add your own data types. As these are facilities unique to Postgres, we support them from Java, with a set of extension API's. Some features within the core of the standard driver actually use these extensions to implement Large Objects, etc. To access some of the extensions, you need to use some extra methods in the org.postgresql.Connection class. In this case, you would need to case the return value of Driver.getConnection(). For example: Connection db = Driver.getConnection(url, username, password); // ... // later on Fastpath fp = ((org.postgresql.Connection)db).getFastpathAPI(); public class Connection extends Object implements Connection java.lang.Object | +----org.postgresql.Connection These are the extra methods used to gain access to PostgreSQL's extensions. Methods defined by java.sql.Connection are not listed. public Fastpath getFastpathAPI() throws SQLException This returns the Fastpath API for the current connection. It is primarily used by the Large Object API. The best way to use this is as follows: import org.postgresql.fastpath.*; ... Fastpath fp = ((org.postgresql.Connection)myconn).getFastpathAPI(); where myconn is an open Connection to PostgreSQL. Fastpath object allowing access to functions on the PostgreSQL backend. SQLException by Fastpath when initializing for first time public LargeObjectManager getLargeObjectAPI() throws SQLException This returns the Large Object API for the current connection. The best way to use this is as follows: import org.postgresql.largeobject.*; ... LargeObjectManager lo = ((org.postgresql.Connection)myconn).getLargeObjectAPI(); where myconn is an open Connection to PostgreSQL. LargeObject object that implements the API SQLException by LargeObject when initializing for first time public void addDataType(String type, String name) This allows client code to add a handler for one of PostgreSQL's more unique data types. Normally, a data type not known by the driver is returned by ResultSet.getObject() as a PGobject instance. This method allows you to write a class that extends PGobject, and tell the driver the type name, and class name to use. The down side to this, is that you must call this method each time a connection is made. The best way to use this is as follows: ... ((org.postgresql.Connection)myconn).addDataType("mytype","my.class.name"); ... where myconn is an open Connection to PostgreSQL. The handling class must extend org.postgresql.util.PGobject. public class Fastpath extends Object java.lang.Object | +----org.postgresql.fastpath.Fastpath Fastpath is an API that exists within the libpq C interface, and allows a client machine to execute a function on the database backend. Most client code will not need to use this method, but it is provided because the Large Object API uses it. To use, you need to import the org.postgresql.fastpath package, using the line: import org.postgresql.fastpath.*; Then, in your code, you need to get a FastPath object: Fastpath fp = ((org.postgresql.Connection)conn).getFastpathAPI(); This will return an instance associated with the database connection that you can use to issue commands. The casing of Connection to org.postgresql.Connection is required, as the getFastpathAPI() is an extension method, not part of JDBC. Once you have a Fastpath instance, you can use the fastpath() methods to execute a backend function. FastpathFastpathArg, LargeObject public Object fastpath(int fnid, boolean resulttype, FastpathArg args[]) throws SQLException Send a function call to the PostgreSQL backend. fnid - Function id resulttype - True if the result is an integer, false for other results args - FastpathArguments to pass to fastpath null if no data, Integer if an integer result, or byte[] otherwise public Object fastpath(String name, boolean resulttype, FastpathArg args[]) throws SQLException Send a function call to the PostgreSQL backend by name. The mapping for the procedure name to function id needs to exist, usually to an earlier call to addfunction(). This is the preferred method to call, as function id's can/may change between versions of the backend. For an example of how this works, refer to org.postgresql.LargeObject name - Function name resulttype - True if the result is an integer, false for other results args - FastpathArguments to pass to fastpath null if no data, Integer if an integer result, or byte[] otherwise LargeObject public int getInteger(String name, FastpathArg args[]) throws SQLException This convenience method assumes that the return value is an Integer name - Function name args - Function arguments integer result SQLException if a database-access error occurs or no result public byte[] getData(String name, FastpathArg args[]) throws SQLException This convenience method assumes that the return value is binary data. name - Function name args - Function arguments byte[] array containing result SQLException if a database-access error occurs or no result public void addFunction(String name, int fnid) This adds a function to our look-up table. User code should use the addFunctions method, which is based upon a query, rather than hard coding the oid. The oid for a function is not guaranteed to remain static, even on different servers of the same version. public void addFunctions(ResultSet rs) throws SQLException This takes a ResultSet containing two columns. Column 1 contains the function name, Column 2 the oid. It reads the entire ResultSet, loading the values into the function table. Remember to close() the ResultSet after calling this! PostgreSQL stores the function id's and their corresponding names in the pg_proc table. To speed things up locally, instead of querying each function from that table when required, a Hashtable is used. Also, only the function's required are entered into this table, keeping connection times as fast as possible. The org.postgresql.LargeObject class performs a query upon its start-up, and passes the returned ResultSet to the addFunctions() method here. Once this has been done, the Large Object API refers to the functions by name. Do not think that manually converting them to the oid's will work. OK, they will for now, but they can change during development (there was some discussion about this for V7.0), so this is implemented to prevent any unwarranted headaches in the future. LargeObjectManager public int getID(String name) throws SQLException This returns the function id associated by its name If addFunction() or addFunctions() have not been called for this name, then an SQLException is thrown. public class FastpathArg extends Object java.lang.Object | +----org.postgresql.fastpath.FastpathArg Each fastpath call requires an array of arguments, the number and type dependent on the function being called. This class implements methods needed to provide this capability. For an example on how to use this, refer to the org.postgresql.LargeObject package. Fastpath, LargeObjectManager, LargeObject public FastpathArg(int value) Constructs an argument that consists of an integer value value - int value to set public FastpathArg(byte bytes[]) Constructs an argument that consists of an array of bytes bytes - array to store public FastpathArg(byte buf[], int off, int len) Constructs an argument that consists of part of a byte array buf source array off offset within array len length of data to include public FastpathArg(String s) Constructs an argument that consists of a String. PostgreSQL has a set of data types that can store geometric features into a table. These include single points, lines, and polygons. We support these types in Java with the org.postgresql.geometric package. It contains classes that extend the org.postgresql.util.PGobject class. Refer to that class for details on how to implement your own data type handlers. Class org.postgresql.geometric.PGbox java.lang.Object | +----org.postgresql.util.PGobject | +----org.postgresql.geometric.PGbox public class PGbox extends PGobject implements Serializable, Cloneable This represents the box data type within PostgreSQL. Variables public PGpoint point[] These are the two corner points of the box. Constructors public PGbox(double x1, double y1, double x2, double y2) Parameters: x1 - first x coordinate y1 - first y coordinate x2 - second x coordinate y2 - second y coordinate public PGbox(PGpoint p1, PGpoint p2) Parameters: p1 - first point p2 - second point public PGbox(String s) throws SQLException Parameters: s - Box definition in PostgreSQL syntax Throws: SQLException if definition is invalid public PGbox() Required constructor Methods public void setValue(String value) throws SQLException This method sets the value of this object. It should be overridden, but still called by subclasses. Parameters: value - a string representation of the value of the object Throws: SQLException thrown if value is invalid for this type Overrides: setValue in class PGobject public boolean equals(Object obj) Parameters: obj - Object to compare with Returns: true if the two boxes are identical Overrides: equals in class PGobject public Object clone() This must be overridden to allow the object to be cloned Overrides: clone in class PGobject public String getValue() Returns: the PGbox in the syntax expected by PostgreSQL Overrides: getValue in class PGobject Class org.postgresql.geometric.PGcircle java.lang.Object | +----org.postgresql.util.PGobject | +----org.postgresql.geometric.PGcircle public class PGcircle extends PGobject implements Serializable, Cloneable This represents PostgreSQL's circle data type, consisting of a point and a radius Variables public PGpoint center This is the center point double radius This is the radius Constructors public PGcircle(double x, double y, double r) Parameters: x - coordinate of center y - coordinate of center r - radius of circle public PGcircle(PGpoint c, double r) Parameters: c - PGpoint describing the circle's center r - radius of circle public PGcircle(String s) throws SQLException Parameters: s - definition of the circle in PostgreSQL's syntax. Throws: SQLException on conversion failure public PGcircle() This constructor is used by the driver. Methods public void setValue(String s) throws SQLException Parameters: s - definition of the circle in PostgreSQL's syntax. Throws: SQLException on conversion failure Overrides: setValue in class PGobject public boolean equals(Object obj) Parameters: obj - Object to compare with Returns: true if the two circles are identical Overrides: equals in class PGobject public Object clone() This must be overridden to allow the object to be cloned Overrides: clone in class PGobject public String getValue() Returns: the PGcircle in the syntax expected by PostgreSQL Overrides: getValue in class PGobject Class org.postgresql.geometric.PGline java.lang.Object | +----org.postgresql.util.PGobject | +----org.postgresql.geometric.PGline public class PGline extends PGobject implements Serializable, Cloneable This implements a line consisting of two points. Currently line is not yet implemented in the backend, but this class ensures that when it's done were ready for it. Variables public PGpoint point[] These are the two points. Constructors public PGline(double x1, double y1, double x2, double y2) Parameters: x1 - coordinate for first point y1 - coordinate for first point x2 - coordinate for second point y2 - coordinate for second point public PGline(PGpoint p1, PGpoint p2) Parameters: p1 - first point p2 - second point public PGline(String s) throws SQLException Parameters: s - definition of the line in PostgreSQL's syntax. Throws: SQLException on conversion failure public PGline() required by the driver Methods public void setValue(String s) throws SQLException Parameters: s - Definition of the line segment in PostgreSQL's syntax Throws: SQLException on conversion failure Overrides: setValue in class PGobject public boolean equals(Object obj) Parameters: obj - Object to compare with Returns: true if the two lines are identical Overrides: equals in class PGobject public Object clone() This must be overridden to allow the object to be cloned Overrides: clone in class PGobject public String getValue() Returns: the PGline in the syntax expected by PostgreSQL Overrides: getValue in class PGobject Class org.postgresql.geometric.PGlseg java.lang.Object | +----org.postgresql.util.PGobject | +----org.postgresql.geometric.PGlseg public class PGlseg extends PGobject implements Serializable, Cloneable This implements a lseg (line segment) consisting of two points Variables public PGpoint point[] These are the two points. Constructors public PGlseg(double x1, double y1, double x2, double y2) Parameters: x1 - coordinate for first point y1 - coordinate for first point x2 - coordinate for second point y2 - coordinate for second point public PGlseg(PGpoint p1, PGpoint p2) Parameters: p1 - first point p2 - second point public PGlseg(String s) throws SQLException Parameters: s - Definition of the line segment in PostgreSQL's syntax. Throws: SQLException on conversion failure public PGlseg() required by the driver Methods public void setValue(String s) throws SQLException Parameters: s - Definition of the line segment in PostgreSQL's syntax Throws: SQLException on conversion failure Overrides: setValue in class PGobject public boolean equals(Object obj) Parameters: obj - Object to compare with Returns: true if the two line segments are identical Overrides: equals in class PGobject public Object clone() This must be overridden to allow the object to be cloned Overrides: clone in class PGobject public String getValue() Returns: the PGlseg in the syntax expected by PostgreSQL Overrides: getValue in class PGobject Class org.postgresql.geometric.PGpath java.lang.Object | +----org.postgresql.util.PGobject | +----org.postgresql.geometric.PGpath public class PGpath extends PGobject implements Serializable, Cloneable This implements a path (a multiply segmented line, which may be closed) Variables public boolean open True if the path is open, false if closed public PGpoint points[] The points defining this path Constructors public PGpath(PGpoint points[], boolean open) Parameters: points - the PGpoints that define the path open - True if the path is open, false if closed public PGpath() Required by the driver public PGpath(String s) throws SQLException Parameters: s - definition of the path in PostgreSQL's syntax. Throws: SQLException on conversion failure Methods public void setValue(String s) throws SQLException Parameters: s - Definition of the path in PostgreSQL's syntax Throws: SQLException on conversion failure Overrides: setValue in class PGobject public boolean equals(Object obj) Parameters: obj - Object to compare with Returns: true if the two pathes are identical Overrides: equals in class PGobject public Object clone() This must be overridden to allow the object to be cloned Overrides: clone in class PGobject public String getValue() This returns the path in the syntax expected by PostgreSQL Overrides: getValue in class PGobject public boolean isOpen() This returns true if the path is open public boolean isClosed() This returns true if the path is closed public void closePath() Marks the path as closed public void openPath() Marks the path as open Class org.postgresql.geometric.PGpoint java.lang.Object | +----org.postgresql.util.PGobject | +----org.postgresql.geometric.PGpoint public class PGpoint extends PGobject implements Serializable, Cloneable This implements a version of java.awt.Point, except it uses double to represent the coordinates. It maps to the point data type in PostgreSQL. Variables public double x The X coordinate of the point public double y The Y coordinate of the point Constructors public PGpoint(double x, double y) Parameters: x - coordinate y - coordinate public PGpoint(String value) throws SQLException This is called mainly from the other geometric types, when a point is embedded within their definition. Parameters: value - Definition of this point in PostgreSQL's syntax public PGpoint() Required by the driver Methods public void setValue(String s) throws SQLException Parameters: s - Definition of this point in PostgreSQL's syntax Throws: SQLException on conversion failure Overrides: setValue in class PGobject public boolean equals(Object obj) Parameters: obj - Object to compare with Returns: true if the two points are identical Overrides: equals in class PGobject public Object clone() This must be overridden to allow the object to be cloned Overrides: clone in class PGobject public String getValue() Returns: the PGpoint in the syntax expected by PostgreSQL Overrides: getValue in class PGobject public void translate(int x, int y) Translate the point with the supplied amount. Parameters: x - integer amount to add on the x axis y - integer amount to add on the y axis public void translate(double x, double y) Translate the point with the supplied amount. Parameters: x - double amount to add on the x axis y - double amount to add on the y axis public void move(int x, int y) Moves the point to the supplied coordinates. Parameters: x - integer coordinate y - integer coordinate public void move(double x, double y) Moves the point to the supplied coordinates. Parameters: x - double coordinate y - double coordinate public void setLocation(int x, int y) Moves the point to the supplied coordinates. refer to java.awt.Point for description of this Parameters: x - integer coordinate y - integer coordinate See Also: Point public void setLocation(Point p) Moves the point to the supplied java.awt.Point refer to java.awt.Point for description of this Parameters: p - Point to move to See Also: Point Class org.postgresql.geometric.PGpolygon java.lang.Object | +----org.postgresql.util.PGobject | +----org.postgresql.geometric.PGpolygon public class PGpolygon extends PGobject implements Serializable, Cloneable This implements the polygon data type within PostgreSQL. Variables public PGpoint points[] The points defining the polygon Constructors public PGpolygon(PGpoint points[]) Creates a polygon using an array of PGpoints Parameters: points - the points defining the polygon public PGpolygon(String s) throws SQLException Parameters: s - definition of the polygon in PostgreSQL's syntax. Throws: SQLException on conversion failure public PGpolygon() Required by the driver Methods public void setValue(String s) throws SQLException Parameters: s - Definition of the polygon in PostgreSQL's syntax Throws: SQLException on conversion failure Overrides: setValue in class PGobject public boolean equals(Object obj) Parameters: obj - Object to compare with Returns: true if the two polygons are identical Overrides: equals in class PGobject public Object clone() This must be overridden to allow the object to be cloned Overrides: clone in class PGobject public String getValue() Returns: the PGpolygon in the syntax expected by PostgreSQL Overrides: getValue in class PGobject Large objects are supported in the standard JDBC specification. However, that interface is limited, and the API provided by PostgreSQL allows for random access to the objects contents, as if it was a local file. The org.postgresql.largeobject package provides to Java the libpq C interface's large object API. It consists of two classes, LargeObjectManager, which deals with creating, opening and deleting large objects, and LargeObject which deals with an individual object. public class LargeObject extends Object java.lang.Object | +----org.postgresql.largeobject.LargeObject This class implements the large object interface to PostgreSQL. It provides the basic methods required to run the interface, plus a pair of methods that provide InputStream and OutputStream classes for this object. Normally, client code would use the getAsciiStream, getBinaryStream, or getUnicodeStream methods in ResultSet, or setAsciiStream, setBinaryStream, or setUnicodeStream methods in PreparedStatement to access Large Objects. However, sometimes lower level access to Large Objects are required, that are not supported by the JDBC specification. Refer to org.postgresql.largeobject.LargeObjectManager on how to gain access to a Large Object, or how to create one. LargeObjectManager public static final int SEEK_SET Indicates a seek from the beginning of a file public static final int SEEK_CUR Indicates a seek from the current position public static final int SEEK_END Indicates a seek from the end of a file public int getOID() Returns the OID of this LargeObject public void close() throws SQLException This method closes the object. You must not call methods in this object after this is called. public byte[] read(int len) throws SQLException Reads some data from the object, and return as a byte[] array public int read(byte buf[], int off, int len) throws SQLException Reads some data from the object into an existing array buf destination array off offset within array len number of bytes to read public void write(byte buf[]) throws SQLException Writes an array to the object public void write(byte buf[], int off, int len) throws SQLException Writes some data from an array to the object buf destination array off offset within array len number of bytes to write public class LargeObjectManager extends Object java.lang.Object | +----org.postgresql.largeobject.LargeObjectManager This class implements the large object interface to PostgreSQL. It provides methods that allow client code to create, open and delete large objects from the database. When opening an object, an instance of org.postgresql.largeobject.LargeObject is returned, and its methods then allow access to the object. This class can only be created by org.postgresql.Connection. To get access to this class, use the following segment of code: import org.postgresql.largeobject.*; Connection conn; LargeObjectManager lobj; // ... code that opens a connection ... lobj = ((org.postgresql.Connection)myconn).getLargeObjectAPI(); Normally, client code would use the getAsciiStream, getBinaryStream, or getUnicodeStream methods in ResultSet, or setAsciiStream, setBinaryStream, or setUnicodeStream methods in PreparedStatement to access Large Objects. However, sometimes lower level access to Large Objects are required, that are not supported by the JDBC specification. Refer to org.postgresql.largeobject.LargeObject on how to manipulate the contents of a Large Object. public static final int WRITE This mode indicates we want to write to an object. public static final int READ This mode indicates we want to read an object. public static final int READWRITE This mode is the default. It indicates we want read and write access to a large object. public LargeObject open(int oid) throws SQLException This opens an existing large object, based on its OID. This method assumes that READ and WRITE access is required (the default). public LargeObject open(int oid, int mode) throws SQLException This opens an existing large object, based on its OID, and allows setting the access mode. public int create() throws SQLException This creates a large object, returning its OID. It defaults to READWRITE for the new object's attributes. public int create(int mode) throws SQLException This creates a large object, returning its OID, and sets the access mode. public void delete(int oid) throws SQLException This deletes a large object. public void unlink(int oid) throws SQLException This deletes a large object. It is identical to the delete method, and is supplied as the C API uses unlink. PostgreSQL is not a normal SQL database. It is more extensible than most other databases, and does support object oriented features that are unique to it. One of the consequences of this, is that you can have one table refer to a row in another table. For example: test=> CREATE TABLE users (username NAME,fullname TEXT); CREATE test=> CREATE TABLE server (servername NAME,adminuser users); CREATE test=> INSERT INTO users VALUES ('peter','Peter Mount'); INSERT 2610132 1 test=> INSERT INTO server VALUES ('maidast',2610132::users); INSERT 2610133 1 test=> SELECT * FROM users; username|fullname --------+-------------- peter |Peter Mount (1 row) test=> SELECT * FROM server; servername|adminuser ----------+--------- maidast | 2610132 (1 row) Okay, the above example shows that we can use a table name as a field, and the row's oid value is stored in that field. What does this have to do with Java? In Java, you can store an object to a Stream as long as it's class implements the java.io.Serializable interface. This process, known as Object Serialization, can be used to store complex objects into the database. Now, under JDBC, you would have to use a Large Object to store them. However, you cannot perform queries on those objects. What the org.postgresql.util.Serialize class does, is provide a means of storing an object as a table, and to retrieve that object from a table. In most cases, you would not need to access this class direct, but you would use the PreparedStatement.setObject() and ResultSet.getObject() methods. Those methods will check the objects class name against the table's in the database. If a match is found, it assumes that the object is a Serialized object, and retrieves it from that table. As it does so, if the object contains other serialized objects, then it recurses down the tree. Sound's complicated? In fact, it's simpler than what I wrote - it's just difficult to explain. The only time you would access this class, is to use the create() methods. These are not used by the driver, but issue one or more CREATE TABLE statements to the database, based on a Java Object or Class that you want to serialize. Oh, one last thing. If your object contains a line like: public int oid; then, when the object is retrieved from the table, it is set to the oid within the table. Then, if the object is modified, and re- serialized, the existing entry is updated. If the oid variable is not present, then when the object is serialized, it is always inserted into the table, and any existing entry in the table is preserved. Setting oid to 0 before serialization, will also cause the object to be inserted. This enables an object to be duplicated in the database. Class org.postgresql.util.Serialize java.lang.Object | +----org.postgresql.util.Serialize public class Serialize extends Object This class uses PostgreSQL's object oriented features to store Java Objects. It does this by mapping a Java Class name to a table in the database. Each entry in this new table then represents a Serialized instance of this class. As each entry has an OID (Object IDentifier), this OID can be included in another table. This is too complex to show here, and will be documented in the main documents in more detail. Constructors public Serialize(org.postgresql.Connection c, String type) throws SQLException This creates an instance that can be used to serialize or deserialize a Java object from a PostgreSQL table. Methods public Object fetch(int oid) throws SQLException This fetches an object from a table, given it's OID Parameters: oid - The oid of the object Returns: Object relating to oid Throws: SQLException on error public int store(Object o) throws SQLException This stores an object into a table, returning it's OID. If the object has an int called OID, and it is > 0, then that value is used for the OID, and the table will be updated. If the value of OID is 0, then a new row will be created, and the value of OID will be set in the object. This enables an object's value in the database to be updateable. If the object has no int called OID, then the object is stored. However if the object is later retrieved, amended and stored again, it's new state will be appended to the table, and will not overwrite the old entries. Parameters: o - Object to store (must implement Serializable) Returns: oid of stored object Throws: SQLException on error public static void create(org.postgresql.Connection con, Object o) throws SQLException This method is not used by the driver, but it creates a table, given a Serializable Java Object. It should be used before serializing any objects. Parameters: c - Connection to database o - Object to base table on Throws: SQLException on error public static void create(org.postgresql.Connection con, Class c) throws SQLException This method is not used by the driver, but it creates a table, given a Serializable Java Object. It should be used before serializing any objects. Parameters: c - Connection to database o - Class to base table on Throws: SQLException on error public static String toPostgreSQL(String name) throws SQLException This converts a Java Class name to a PostgreSQL table, by replacing . with _ Because of this, a Class name may not have _ in the name. Another limitation, is that the entire class name (including packages) cannot be longer than 31 characters (a limit forced by PostgreSQL). Parameters: name - Class name Returns: PostgreSQL table name Throws: SQLException on error public static String toClassName(String name) throws SQLException This converts a PostgreSQL table to a Java Class name, by replacing _ with . Parameters: name - PostgreSQL table name Returns: Class name Throws: SQLException on error Utility Classes The org.postgresql.util package contains classes used by the internals of the main driver, and the other extensions. Class org.postgresql.util.PGmoney java.lang.Object | +----org.postgresql.util.PGobject | +----org.postgresql.util.PGmoney public class PGmoney extends PGobject implements Serializable, Cloneable This implements a class that handles the PostgreSQL money type Variables public double val The value of the field Constructors public PGmoney(double value) Parameters: value - of field public PGmoney(String value) throws SQLException Create a money. Parameters: value - Definition of this money in PostgreSQL's syntax public PGmoney() Required by the driver Methods public void setValue(String s) throws SQLException Parameters: s - Definition of this money in PostgreSQL's syntax Throws: SQLException on conversion failure Overrides: setValue in class PGobject public boolean equals(Object obj) Parameters: obj - Object to compare with Returns: true if the two moneys are identical Overrides: equals in class PGobject public Object clone() This must be overridden to allow the object to be cloned Overrides: clone in class PGobject public String getValue() Returns: the PGmoney in the syntax expected by PostgreSQL Overrides: getValue in class PGobject Class org.postgresql.util.PGobject java.lang.Object | +----org.postgresql.util.PGobject public class PGobject extends Object implements Serializable, Cloneable This class is used to describe data types that are unknown by JDBC Standard. A call to org.postgresql.Connection permits a class that extends this class to be associated with a named type. This is how the org.postgresql.geometric package operates. ResultSet.getObject() will return this class for any type that is not recognized on having it's own handler. Because of this, any PostgreSQL data type is supported. Constructors public PGobject() This is called by org.postgresql.Connection.getObject() to create the object. Methods public final void setType(String type) This method sets the type of this object. It should not be extended by subclasses, hence its final Parameters: type - a string describing the type of the object public void setValue(String value) throws SQLException This method sets the value of this object. It must be overridden. Parameters: value - a string representation of the value of the object Throws: SQLException thrown if value is invalid for this type public final String getType() As this cannot change during the life of the object, it's final. Returns: the type name of this object public String getValue() This must be overridden, to return the value of the object, in the form required by PostgreSQL. Returns: the value of this object public boolean equals(Object obj) This must be overridden to allow comparisons of objects Parameters: obj - Object to compare with Returns: true if the two objects are identical Overrides: equals in class Object public Object clone() This must be overridden to allow the object to be cloned Overrides: clone in class Object public String toString() This is defined here, so user code need not override it. Returns: the value of this object, in the syntax expected by PostgreSQL Overrides: toString in class Object Class org.postgresql.util.PGtokenizer java.lang.Object | +----org.postgresql.util.PGtokenizer public class PGtokenizer extends Object This class is used to tokenize the text output of PostgreSQL. We could have used StringTokenizer to do this, however, we needed to handle nesting of '(' ')' '[' ']' '<' and '>' as these are used by the geometric data types. It's mainly used by the geometric classes, but is useful in parsing any output from custom data types output from PostgreSQL. See Also: PGbox, PGcircle, PGlseg, PGpath, PGpoint, PGpolygon Constructors public PGtokenizer(String string, char delim) Create a tokenizer. Parameters: string - containing tokens delim - single character to split the tokens Methods public int tokenize(String string, char delim) This resets this tokenizer with a new string and/or delimiter. Parameters: string - containing tokens delim - single character to split the tokens public int getSize() Returns: the number of tokens available public String getToken(int n) Parameters: n - Token number ( 0 ... getSize()-1 ) Returns: The token value public PGtokenizer tokenizeToken(int n, char delim) This returns a new tokenizer based on one of our tokens. The geometric data types use this to process nested tokens (usually PGpoint). Parameters: n - Token number ( 0 ... getSize()-1 ) delim - The delimiter to use Returns: A new instance of PGtokenizer based on the token public static String remove(String s, String l, String t) This removes the lead/trailing strings from a string Parameters: s - Source string l - Leading string to remove t - Trailing string to remove Returns: String without the lead/trailing strings public void remove(String l, String t) This removes the lead/trailing strings from all tokens Parameters: l - Leading string to remove t - Trailing string to remove public static String removePara(String s) Removes ( and ) from the beginning and end of a string Parameters: s - String to remove from Returns: String without the ( or ) public void removePara() Removes ( and ) from the beginning and end of all tokens public static String removeBox(String s) Removes [ and ] from the beginning and end of a string Parameters: s - String to remove from Returns: String without the [ or ] public void removeBox() Removes [ and ] from the beginning and end of all tokens public static String removeAngle(String s) Removes < and > from the beginning and end of a string Parameters: s - String to remove from Returns: String without the < or > public void removeAngle() Removes < and > from the beginning and end of all tokens Class org.postgresql.util.Serialize This was documented earlier under Object Serialization. Class org.postgresql.util.UnixCrypt java.lang.Object | +----org.postgresql.util.UnixCrypt public class UnixCrypt extends Object This class provides us with the ability to encrypt passwords when sent over the network stream Contains static methods to encrypt and compare passwords with Unix encrypted passwords. See John Dumas's Java Crypt page for the original source. (Invalid URL) http://www.zeh.com/local/jfd/crypt.html Methods public static final String crypt(String salt, String original) Encrypt a password given the clear-text password and a salt. Parameters: salt - A two-character string representing the salt used to iterate the encryption engine in lots of different ways. If you are generating a new encryption then this value should be randomized. original - The password to be encrypted. Returns: A string consisting of the 2-character salt followed by the encrypted password. public static final String crypt(String original) Encrypt a password given the clear-text password. This method generates a random salt using the 'java.util.Random' class. Parameters: original - The password to be encrypted. Returns: A string consisting of the 2-character salt followed by the encrypted password. public static final boolean matches(String encryptedPassword, String enteredPassword) Check that enteredPassword encrypts to encryptedPassword. Parameters: encryptedPassword - The encryptedPassword. The first two characters are assumed to be the salt. This string would be the same as one found in a Unix /etc/passwd file. enteredPassword - The password as entered by the user (or otherwise acquired). Returns: true if the password should be considered correct. A problem with many JDBC drivers is that only one thread can use a Connection at any one time -- otherwise a thread could send a query while another one is receiving results, and this would be a bad thing for the database engine. PostgreSQL 6.4 brought thread safety to the entire driver. (Standard JDBC was thread safe in 6.3, but the Fastpath API was not.) Consequently, if your application uses multiple threads then you do not have to worry about complex algorithms to ensure that only one uses the database at any time. If a thread attempts to use the connection while another one is using it, it will wait until the other thread has finished its current operation. If it is a regular SQL statement, then the operation consists of sending the statement and retrieving any ResultSet (in full). If it is a Fastpath call (e.g., reading a block from a LargeObject) then it is the time to send and retrieve that block. This is fine for applications and applets but can cause a performance problem with servlets. With servlets you can have a heavy load on the connection. If you have several threads performing queries then each but one will pause, which may not be what you are after. To solve this, you would be advised to create a pool of connections. When ever a thread needs to use the database, it asks a manager class for a Connection. The manager hands a free connection to the thread and marks it as busy. If a free connection is not available, it opens one. Once the thread has finished with it, it returns it to the manager who can then either close it or add it to the pool. The manager would also check that the connection is still alive and remove it from the pool if it is dead. So, with servlets, it is up to you to use either a single connection, or a pool. The plus side for a pool is that threads will not be hit by the bottle neck caused by a single network connection. The down side is that it increases the load on the server, as a backend process is created for each Connection. It is up to you and your applications requirements. If you have not yet read it, I'd advise you read the JDBC API Documentation (supplied with Sun's JDK), and the JDBC Specification. Both are available from http://java.sun.com/products/jdbc/index.html. http://jdbc.postgresql.org contains updated information not included in this document, and also includes precompiled drivers.