Java Fundamentals 4andrei.ase.ro/assets/codeclub/ep3.pdfJDBC vs. ODBC –JDBC is not the first try at standardized data access –ODBC (open database connectivity) is based on native

Java Fundamentals 4

JDBC

What is JDBC?

– Java Database Connectivity

– JDBC is a standard Java API implementing access to tabular data (especially relational databases)

Java Application

Driver Manager

JDBC Driver

Why use JDBC?

– JDBC provides standardized access todatabases

– JDBC allows you to change the databasewithout changing the Java code

– JDBC is an intermediary data access layer

– JDBC provides parametrized queries, usefulin batch operations and avoiding SQLinjection

JDBC vs. ODBC

– JDBC is not the first try at standardizeddata access

– ODBC (open database connectivity) isbased on native implementations

– JDBC is independent from the OSimplementation, unlike ODBC. When usingODBC support, you are dependent on OSfeatures

– The JDBC-ODBC bridge is a special driverfor ODBC access via JDBC

JDBC Drivers

– Drivers are implementations handlingactual communication to the database

– An application can load (register) multipledrivers

– Drivers are registered with DriverManager

– Only one instance of a driver is loaded

Types of JDBC Drivers

– Type 1 - JDBC-ODBC bridge

• translates queries into ODBC queries

– Type 2 - Native API driver

• uses the native database client

Types of JDBC Drivers

– Type 3 - Network protocol driver

• uses a middle tier

• requests are done through an applicationserver

– Type 4 - Native protocol driver

• pure java driver

• communicates with the database via it ownprotocol, usually via sockets

Typical JDBC application

Register a JDBC driver

Open a connection

Create a statement

Execute the

statement

Process results

Close the connection

Loading a driver

– Ways of loading a driver

• Through the command line

– Via a properties file

– Through Class.forName

• the driver is loaded via a static block in theclass

– Via an object of the driver class

Loading a driver - examples

1. //driver loaded from command line

2. java -Djdbc.drivers=com.mysql.jdbc.Driver Test

3. //driver loaded through a static block

4. Class.forName("com.mysql.jdbc.Driver");

5. //driver loaded through the driver manager

6. Driver driver = new com.mysql.jdbc.Driver();

7. DriverManager.registerDriver(driver);

Getting a connection

– DriverManager intermediates gettingconnections from the database

– Connections are requested via an URL

– If the connection requires a username andpassword, they can be supplied viaoverloaded implementations ofgetConnection

Getting a connection

1. String url = jdbc:mysql://localhost:3306/acmeinc";

2. String user = "root";

3. String password = "welcome123";

4. conn = DriverManager.getConnection(url, user, password);

Closing a connection

– Connections are rare resources and shouldbe closed

– Connections should be closed on finally

– Connection state can be verified with theisClosed method

Closing a connection

1. //close a connection if it was opened

2. finally{

3. if (conn!=null)

4. conn.close();

5. }

6. //verify if a connection is opened

7. if(!conn.isClosed()){

8. //use the connection

9. }

JDBC URL's

– General form:1. jdbc:<database_protocol>:<identifier>

– JDBC URL's are, however, not standard

– Subprotocol is the protocol used by thedatabase

– Identifier is usually a catalog identifier1. jdbc:mysql://localhost:3306/acmeinc

– Identifier can also contain databasespecific properties1. jdbc:odbc:testdb;UID=user;PWD=pass

Connection pools

– Connection pools allow for reuse ofconnections

– Connections are not closed manually

– Usually connection pools are implementedin application containers

Types of statements

– Statement objects execute simple SQL

• vulnerable to injection

• should only be used when no parameters arepassed

– PreparedStatement objects representparametrized statements

• forces parameter type

• avoids injection attacks

Types of statements

– CallableStatement objects are used forinteraction with stored procedures

• allows for retrieval of parameter values

• supports in, out and dual purpose parameters

• forces parameter type

Statements

– Statement objects execute a queryrepresented by a String

– If the String is parametrized as below, aseries of vulnerabilities emerge

• String query = "SELECT * FROMusers

• WHERE user = " + userString;

– Parametrized statements should bepreferred in all cases where a Statementexecutes a command containing user input

Using a Statement

– Creating a Statement1. Statement stmt = null;

2. stmt = conn.createStatement();

– Executing a query1. ResultSet rs =

2. stmt.executeQuery("SELECT * FROM users");

– Executing an update1. int result = stmt.executeUpdate("INSERT INTO

2. employees(idemployees, full_name, type)

3. VALUES (3,'Orr','mistery') ");

Prepared Statements

– Prepared statements representparametrized queries

– Prepared statements are usually compiledon the database server (though it dependson the actual driver implementation)

– Prepared statements are faster forrepeated query, but they carry someoverhead for single queries

Prepared Statements

– Prepared statements force parameter typeand make it impossible to executeparameter content

– In order to avoid injection attacks,prepared statements should always beused where parameters come directly orindirectly from an user

Using a PreparedStatement

– Creating a prepared statement1. PreparedStatement stmt = null;

2. stmt = conn.prepareStatement("UPDATE

3. employees SET type = ? WHERE idemployees

4. = ?");

– Setting parameter values is donepositionally (1 based)1. stmt.setString(1, "deserted");

2. stmt.setInt(2, 3);

Using a PreparedStatement

– Executing the statement(same types ofresults as Statement objects)1. int result = stmt.executeUpdate();

Returned values

– Statements and prepared statementsreturn either an integer value or a resultset

• Integer values usually have the significance of"number of rows affected"

• Result sets are obtained by executing a selectstatement

Query results

– Queries (Selects) return result sets

– Result sets are represented by ResultSetobjects

• ResultSets are not serializable because theycontain a reference to a Connection object

– RowSets are another way of retrievingresults

• RowSets do not have a standardimplementation

• RowSets are serializable

Result sets

– Results of selects are represented asResultSet objects

– ResultSet objects are iterable collections oftuples

Result sets

– Result sets do however have a lot morehidden complexity, as they actuallyrepresent cursors on the database

• one consequence is that, provided the driverpermits it, result sets can be accessedpositionally

• result sets can also be treated as updateable(updating the result updates the underlyingdata)

Retrieving results

– Getting a result set from a statement1. ResultSet rs =

2. stmt.executeQuery("SELECT * FROM

3. employees");

– Getting a result set from a preparedstatement1. ResultSet rs =

2. stmt.executeQuery("SELECT * FROM

3. employees");

Iterating through results

– Result sets have a next() method whichreturns

• false if no more results are available

• true if more results are available; in this situation, the method also advances to the next result

– Iterating through the result set1. while(rs.next()){

2. //process row

3. }

Retrieving simple types

– In order to retrieve values from a resultset, SQL types are mapped to data types

– Columns can be retrieved by position(1based) or name1. int id = rs.getInt(1);

2. String name = rs.getString("full_name");

Retrieving function results

– Function and their results are DBMSspecific

– Function results are retrieved in columns,just like regular columns

– While is still used as a check for the factthat the function returned 1 value(insteadon none) even if the function returns anaggregate value

Retrieving function results

– If the function is a map function (receivesa list of values and returns a list of values),there will be as many values as tuples

– While is still used as a check for the factthat the function returned 1 value(insteadon none)1. rs = stmt.executeQuery("SELECT COUNT(*) FROM

2. employees");

3. while(rs.next()){

4. int i = rs.getInt(1);

5. }

Retrieving function values

– If the function is a reduce function(afunction that receives a list of values andproduces one aggregate result), there willbe only one row, containing one value forthe function1. rs = stmt.executeQuery("SELECT

2. UPPER(full_name) FROM employees");

3. while(rs.next()){

4. System.out.println("Apparent name " +

5. rs.getString(1));

6. }

Callable statements

– Callable statement facilitate the interactionwith stored procedures

– Usually stored procedures return valuesreadable from variables defined in theDBMS itself

– Such values are not extractable from aresult set

– Callable statements support input, outputand dual purpose parameters

Using a callable statement

– Creating a callable statement1. CallableStatement stmt = null;

2. stmt = conn.prepareCall("CALL

3. test_proc(?,?,?)");

– Setting input parameters1. stmt.setInt(1, 10);

– Registering output parameters1. stmt.registerOutParameter(3, Types.INTEGER);

Using a callable statement

– Handling input/output parameters (INOUTparameters have to be both set andregistered )1. stmt.setInt(2, 20);

2. stmt.registerOutParameter(2, Types.INTEGER);

– Executing the callable statement1. stmt.execute();

Using a callable statement

– Retrieving results(results are retrievedfrom output and dual purpose parameters)1. int y = stmt.getInt("y");

2. int z = stmt.getInt("z");

Retrieving result metadata

– Sometimes the database schema must bediscovered at runtime

– ResultSetMetaData can be obtained from aResultSet1. ResultSet rs = null;

2. ResultSetMetaData rsmd = null;

3. //retrieve result set

4. rsmd = rs.getMetaData();

Retrieving result metadata

– Getting the number of columns1. int nrCol = rsmd.getColumnCount();

– Getting the name of a column(1 based)1. String colName = rsmd.getColumnName(1);

– Getting the type of a column (the typecorresponds to the enumerationjava.sql.Types)1. int type = rsmd.getColumnType(i);

Retrieving result metadata

– Getting the name of the catalog(database); can be applied to any column1. String catalog = rsmd.getCatalogName(1);

– Some drivers do not implement bothgetting the table and the schema of a table

– Getting the name of the schema (table);can be applied to any column1. String table = rsmd.getTableName(1);

2. String schema = rsmd.getSchemaName(1);

Handling large objects

– Databases usually support the storage oflarge objects

– The typical associated types are

• CLOB - character large objects with no Javaequivalent

• BLOB - binary large objects with no Javaequivalent

Handling large objects

– Two approaches are possible

• using streams

• using setXXX methods

• on MySql, setBinaryStream should be used

– Connection limits have to be taken intoaccount when dealing with multimegabytecontent

• in this situation streams are the only solution

Inserting a BLOB

1. stmt = conn.prepareStatement("INSERT INTO blobs

2. VALUES (?,?)");

3. stmt.setInt(1, 2);

4. File file = new File("test.jpeg");

5. FileInputStream input = new FileInputStream(file);

6. stmt.setBinaryStream(2, input,file.length());

7. int result = stmt.executeUpdate();

Recovering a BLOB

1. stmt = conn.prepareStatement("SELECT blobscol FROM

2. blobs");

3. ResultSet rs = stmt.executeQuery();

4. while (rs.next()){

5. file = new File("out.jpg");

6. FileOutputStream output = new

7. FileOutputStream(file);

8. InputStream reader = rs.getBinaryStream(1);

9. byte[] b = new byte[1];

10. while(reader.read(b)>0) output.write(b);

11. output.close();

12. }

Batch processing

– PreparedStatements are much moreefficient when reused

– Reuse is facilitated by a batchingmechanism

– Through batching, prepared statementswith the same template are grouped

– A batch of statements produces an array ofresults (one for each statement)

Using batches

– Preparing the statement1. stmt = conn.prepareStatement("INSERT INTO

2. simpletable VALUES (?,?)");

– Setting parameters and adding each batch1. stmt.setInt(1, 1);

2. stmt.setInt(2, 3);

3. stmt.addBatch();

4. stmt.setInt(1, 2);

5. stmt.setInt(2, 4);

6. stmt.addBatch();

Using batches

– Executing the batch1. int[] results = stmt.executeBatch();

– Accessing the results1. for (int i = 0; i < results.length; i++) {

2. int j = results[i];

3. }

Managing transactions

– Usually transactions are autocommited(each operation is a transaction)

– In order to avoid it, autocommit must beset to false1. conn.setAutoCommit(false);

Managing transactions

– In order to revert to the anterior state, theconnection can be rolled back1. conn.rollback();

– In any event for the data to be modified,the connection must be commited1. conn.commit();

More transactions

– More complex transactions can beimplemented by defining varioussavepoints1. Savepoint save = conn.setSavepoint();

– A connection can be rolled back to anydefined savepoint1. conn.rollback(save);

– After they are no longer needed,savepoints must be released1. conn.releaseSavepoint(save);

SQL escape syntax

– When writing SQL queries as simplestrings, the syntax must be that of theparticular DBMS

– JDBC has a mechanism through whichmore general, DBMS independent queriescan be written

– Sql escape syntax defines general functionswhich are then translated by the driver

– Escaped sequences are written between {}

Escape syntax examples

– Calling a procedure1. "{call my_stored_procedure ?, ?}"

– Defining a date1. "{d ‘2011-11-10'}"

– Calculating a function1. "{fn ucase('alfa')}"

Autogenerated keys

– Autogenerated keys are a common featureof DBMS's

– The reason is that it is impossible to get anunique key without multiple queries

– A few restrictions must be respected

• columns must be specified by name in theinsert statement

• the return generated keys option must be usedto configure the statement

Autogenerated keys

– Declaring a prepared statement capable ofretrieving keys1. stmt = conn.prepareStatement("INSERT INTO

2. simpleauto (content) VALUES

3. (?)",Statement.RETURN_GENERATED_KEYS);

– Retrieving the keys from a result set (eachkey is a row in the result set)1. ResultSet rs = stmt.getGeneratedKeys();

2. while (rs.next()) {

3. System.out.println(rs.getString(1));

4. }

Other topics

– Updateable results

– Result sets and cursors

– Retrieving database metadata

– Setting result set properties

– Setting connection properties

– Handling exceptions and warnings