JavaDBF is a Java library for reading and writing XBase files. There are plenty of legacy applications around with .dbf as their primary storage format. JavaDBF was initially written for data transfer with such applications.
JavaDBF is LGPL
https://www.gnu.org/licenses/lgpl.txt
Till late 90s dBase and its cousins were the most preferred database platform for small and even medium enterprise applications. They required low hardware configurations and were cheaper to develop. Eventually more capable desktop databases like Microsoft Access came into picture, DBF file format still remains one of the simplest way to store and transfer data.
JavaDBF 1.0.0 is the first release considered stable. It has a lot of improvements and is almost compatible with old code. But there are some small changes that may break your code:
- Numeric and Double types (N and F) are now returned as BigDecimal instead of Double.
- DBFException is now a RuntimeException, not need to explictitly catch or rethrow, but yo can if you wish.
- Character data is now trimmed tailing white spaces. You can disable it calling reader.setTrimRightSpaces(false)
- Support for dbase Level 7 databases.
- Support for new field types.
- Support for memo text and binary types (M, B, G, P) via external file (fpt or dbt)
- Support for varchar (V) fields
- Support for double (O) fields
- Support for double (B) fields
- Support for long character (up to 65534 bytes) fields
Obtain the latest version of JavaDBF from release page at github. Download the jar file and put it in your $CLASSPATH variable. You are ready to go.
If you are using Maven, you can add JavaDBF to your project using this dependency in your pom.xml
<dependency>
<groupId>com.github.albfernandez</groupId>
<artifactId>javadbf</artifactId>
<version>1.13.1</version>
</dependency>
JavaDBF has a simple API of its own and it does not implement the JDBC API. It is designed this way because JavaDBF is not intended to support full-blown RDBMS-style database interaction. And you are not supposed to use it like a back-end; it just doesn't work that way. Also, JavaDBF is not designed to be thread-safe; keep that in mind when you design multithreaded applications.
JavaDBF comes in the package com.linuxense.javadbf. Import that package in your Java code. Following examples will familiarise you with its APIs.
JavaDBF supports almost all XBase data types. While reading, those types are interpretted as appropriate Java types. Following tables shows the mapping scheme.
XBase Type | XBase Symbol | Java Type used in JavaDBF |
---|---|---|
Character | C | java.lang.String |
Numeric | N | java.math.BigDecimal |
Floating Point | F | java.math.BigDecimal |
Logical | L | java.lang.Boolean |
Date | D | java.util.Date |
FoxPro Type | Symbol | Java Type used in JavaDBF |
---|---|---|
Currency | Y | java.math.BigDecimal |
Long | I | java.lang.Integer |
Date Type | T | java.util.Date |
Timestamp | @ | java.util.Date |
AutoIncrement | + | java.lang.Integer |
Memo | M | java.lang.String or byte[] |
Binary | B | byte[] or java.lang.Double |
Blob | W | byte[] |
General | G | byte[] |
Picture | P | byte[] |
VarBinary | Q | byte[] |
Varchar | V | java.lang.String |
Double | O | java.lang.Double |
To read a DBF file, JavaDBF provides a DBFReader class. Following is a ready-to-compile, self-explanatory program describing almost all feature of the DBFReader class. Copy/paste this listing and compile it. Keep a .dbf file handy to pass to this program as its argument.
import java.io.*;
import com.linuxense.javadbf.*;
public class JavaDBFReaderTest {
public static void main(String args[]) {
DBFReader reader = null;
try {
// create a DBFReader object
reader = new DBFReader(new FileInputStream(args[0]));
// get the field count if you want for some reasons like the following
int numberOfFields = reader.getFieldCount();
// use this count to fetch all field information
// if required
for (int i = 0; i < numberOfFields; i++) {
DBFField field = reader.getField(i);
// do something with it if you want
// refer the JavaDoc API reference for more details
//
System.out.println(field.getName());
}
// Now, lets us start reading the rows
Object[] rowObjects;
while ((rowObjects = reader.nextRecord()) != null) {
for (int i = 0; i < rowObjects.length; i++) {
System.out.println(rowObjects[i]);
}
}
// By now, we have iterated through all of the rows
} catch (DBFException e) {
e.printStackTrace();
} catch (IOException e) {
e.printStackTrace();
}
finally {
DBFUtils.close(reader);
}
}
}
From JavaDBF 1.7.0 you can use field names to access data. You must use nextRow instead nextRecord
import java.io.*;
import com.linuxense.javadbf.*;
public class JavaDBFReaderWithFieldNamesTest {
public static void main(String args[]) {
DBFReader reader = null;
try {
// create a DBFReader object
reader = new DBFReader(new FileInputStream(args[0]));
// get the field count if you want for some reasons like the following
int numberOfFields = reader.getFieldCount();
// use this count to fetch all field information
// if required
for (int i = 0; i < numberOfFields; i++) {
DBFField field = reader.getField(i);
// do something with it if you want
// refer the JavaDoc API reference for more details
//
System.out.println(field.getName());
}
// Now, lets us start reading the rows
DBFRow row;
while ((row = reader.nextRow()) != null) {
System.out.println(row.getString("PHONE"));
}
// By now, we have iterated through all of the rows
} catch (DBFException e) {
e.printStackTrace();
} catch (IOException e) {
e.printStackTrace();
}
finally {
DBFUtils.close(reader);
}
}
}
You can specify memo file to read Memo fields from. If you don't specify this file this fields will be filled with null
import java.io.*;
import com.linuxense.javadbf.*;
public class JavaDBFReaderMemoTest {
public static void main(String args[]) {
DBFReader reader = null;
try {
// create a DBFReader object
reader = new DBFReader(new FileInputStream(args[0]));
reader.setMemoFile(new File("memo.dbt"));
// do whatever you want with the data
} catch (DBFException e) {
e.printStackTrace();
} catch (IOException e) {
e.printStackTrace();
} finally {
DBFUtils.close(reader);
}
}
}
The class complementary to DBFReader is the DBFWriter. While creating a .dbf data file you will have to deal with two aspects:
- define the fields and
- populate data.
As mentioned above a dbf field is represented by the class DBFField. First, let us familiarise this class.
Create an object of DBFField class:
DBFField field = new DBFField();
field.setName("emp_name"); // give a name to the field
field.setType(DBFDataType.CHARACTER); // and set its type
field.setLength(25); // and length of the field
This is, now, a complete DBFField Object ready to use. We have to create as many DBFField Objects as we want to be in the .dbf file. The DBFWriter class accept DBFField in an array. Now, let's move on to the next step of populating data.
A DBFWriter is used for creating a .dbf file. First lets create a DBFWriter object by calling its constructor and then set the fields created (as explained above) by calling the setFields method.
OutputStream os = // wathever output you want to use
DBFWriter writer = new DBFWriter(os);
// fields is a non-empty array of DBFField objects
writer.setFields(fields);
Now, the DBFWriter Object is ready to be populated. The method for adding data to the DBFWriter is addRecord and it takes an Object array as its argument. This Object array is supposed contain values for the fields added with one-to-one correspondence with the fields set.
Following is a complete program explaining all the steps described above:
import com.linuxense.javadbf.*;
import java.io.*;
public class JavaDBFWriterTest {
public static void main(String args[]) throws IOException {
// let us create field definitions first
// we will go for 3 fields
DBFField[] fields = new DBFField[3];
fields[0] = new DBFField();
fields[0].setName("emp_code");
fields[0].setType(DBFDataType.CHARACTER);
fields[0].setLength(10);
fields[1] = new DBFField();
fields[1].setName("emp_name");
fields[1].setType(DBFDataType.CHARACTER);
fields[1].setLength(20);
fields[2] = new DBFField();
fields[2].setName("salary");
fields[2].setType(DBFDataType.NUMERIC);
fields[2].setLength(12);
fields[2].setDecimalCount(2);
DBFWriter writer = new DBFWriter(new FileOutputStream(args[0]));
writer.setFields(fields);
// now populate DBFWriter
Object rowData[] = new Object[3];
rowData[0] = "1000";
rowData[1] = "John";
rowData[2] = new Double(5000.00);
writer.addRecord(rowData);
rowData = new Object[3];
rowData[0] = "1001";
rowData[1] = "Lalit";
rowData[2] = new Double(3400.00);
writer.addRecord(rowData);
rowData = new Object[3];
rowData[0] = "1002";
rowData[1] = "Rohit";
rowData[2] = new Double(7350.00);
writer.addRecord(rowData);
// write to file
writer.close();
}
}
Keep in mind that till the close method is called, all the added data will be kept in memory. So, if you are planning to write huge amount of data make sure that it will be safely held in memory till it is written to disk and the DBFWriter object is garbage-collected. Read the Sync Mode section to know how JavaDBF to use a special feature of JavaDBF to overcome this.
This is useful when JavaDBF is used to create a DBF with very large number of records. In this mode, instead of keeping records in memory for writing them once for all, records are written to file as addRecord() is called. Here is how to write in Sync Mode.
Create DBFWriter instance by passing a File object which represents a new/non-existent or empty file. And you are done! But, as in the normal mode, remember to call close() when have added all the records. This will help JavaDBF to write the meta data with correct values. Here is a sample code:
import com.linuxense.javadbf.*;
import java.io.*;
public class DBFWriterTest {
public static void main(String args[]) throws IOException {
// ...
DBFWriter writer = new DBFWriter(new File("/path/to/a/new/file"));
// this DBFWriter object is now in Sync Mode
// set fields
writer.setFields(fields);
// add the data
writer.addRecord(rowData);
// close
writer.close();
// ...
}
}
From version 0.4.0 onwards JavaDBF supports appending of records to an existing DBF file. Use the same constructor used in Sync Mode to achieve this. But here the File object passed to the construction should represent the DBF file to which records are to be appended.
It is illegal to call setFields in DBFWriter object created for appending. Here also it is required to call the close() method after adding all the records.
import com.linuxense.javadbf.*;
import java.io.*;
public class DBFWriterTest {
public static void main(String args[]) throws IOException {
// ...
DBFWriter writer = new DBFWriter(new File("/path/to/an/existing/dbfile"));
// this DBFWriter object is now in Syc Mode
// add the data (no setFields, because it's an existing file)
writer.addRecord(rowData);
// close
writer.close();
// ...
}
}
Clone the repository or download de tar file from releases page on github, then run the Maven command:
git clone https://github.com/albfernandez/javadbf.git
cd javadbf
mvn clean package
The result file is target/javadbf-x.xx.x.jar
https://en.wikipedia.org/wiki/.dbf
http://www.clicketyclick.dk/databases/xbase/format/index.html
http://www.dbase.com/Knowledgebase/INT/db7_file_fmt.htm
http://msdn.microsoft.com/en-us/library/st4a0s68%28VS.80%29.aspx