PostgreSQL JDBC driver
This is a simple readme describing how to compile and use the PostgreSQL JDBC driver.
Downloading pre-built drivers
Most people do not need to compile PgJDBC. You can download prebuilt versions of the driver from the Postgresql JDBC site or using your chosen dependency management tool:
<dependency> <groupId>org.postgresql</groupId> <artifactId>postgresql</artifactId> <version>9.4.1208</version> <!-- Java 8 --> <version>9.4.1208.jre7</version> <!-- Java 7 --> <version>9.4.1208.jre6</version> <!-- Java 6 --> </dependency>
<dependency org="org.postgresql" name="postgresql" rev="9.4.1208"/>
<dependency org="org.postgresql" name="postgresql" rev="9.4.1208.jre7"/>
<dependency org="org.postgresql" name="postgresql" rev="9.4.1208.jre6"/>
Snapshot builds (builds from
master branch) are deployed to Maven Central, so you can test current development version via
<dependency> <groupId>org.postgresql</groupId> <artifactId>postgresql</artifactId> <version>9.4.1209-SNAPSHOT</version> <!-- Java 8 --> <version>9.4.1209.jre7-SNAPSHOT</version> <!-- Java 7 --> <version>9.4.1209.jre6-SNAPSHOT</version> <!-- Java 6 --> </dependency>
In order to build the source code for PgJDBC you will need the following tools:
- A git client
- A recent version of Maven (3.x)
- A JDK for the JDBC version you'd like to build (JDK6 for JDBC 4, JDK7 for JDBC 4.1 or JDK8 for JDBC 4.2)
- A running PostgreSQL instance
Additionally, in order to update translations (not typical), you will need the following additional tools:
- the gettext package, which contains the commands "msgfmt", "msgmerge", and "xgettext"
Checking out the source code
The PgJDBC project uses git for version control. You can check out the current code by running:
git clone https://github.com/pgjdbc/pgjdbc.git
This will create a pgjdbc directory containing the checked-out source code.
In order do build jre7 or jre6 compatible versions, check out those repositories under
cd pgjdbc # <-- that is pgjdbc/pgjdbc.git clone git clone https://github.com/pgjdbc/pgjdbc-jre7.git git clone https://github.com/pgjdbc/pgjdbc-jre6.git
Note: all the source code is stored in
pgjdbc.git repository, so just
pgjdbc.git is sufficient for development.
Compiling with Maven on the command line
After checking out the code you can compile and test the PgJDBC driver by running the following on a command line:
Note: if you want to skip test execution, issue
mvn package -DskipTests.
Note: in certain cases, proper build requires cleaning the results of previous one.
For instance, if you remove a
.java file, then clean is required to remove the relevant
In such cases, use
mvn clean or
mvn clean package.
PgJDBC doesn't natively support building from IDEs like Eclipse, NetBeans or
IntelliJ. However you can use the tools Maven support from within the IDE if you wish.
You can use regular IDE tools to develop, execute tests, etc, however if you want to build final
artifacts you should use
After running the build , and build a .jar file (Java ARchive) depending on the version of java and which release you have the jar will be named postgresql-...jre.jar. Where major,minor are the postgreSQL major,minor version numbers. release is the jdbc release number. N is the version of the JDBC API which corresponds to the version of Java used to compile the driver.
The target directory will contain the driver jar.
If you need source code, documentation and runtime dependencies use
mvn package -P release-artifacts.
NOTE: default build produces Java 8 (JDBC 4.2) driver (in
If you need a version for older Java, configure
Here's sample configuration for Mac OS:
<?xml version="1.0" encoding="UTF8"?> <toolchains> <!-- JDK toolchains --> <toolchain> <type>jdk</type> <provides> <version>1.6</version> <vendor>oracle</vendor> </provides> <configuration> <jdkHome>/Library/Java/JavaVirtualMachines/1.6.0.jdk/Contents/Home</jdkHome> </configuration> </toolchain> <toolchain> <type>jdk</type> <provides> <version>1.7</version> <vendor>oracle</vendor> </provides> <configuration> <jdkHome>/Library/Java/JavaVirtualMachines/jdk1.7.0_55.jdk/Contents/Home</jdkHome> </configuration> </toolchain> <toolchain> <type>jdk</type> <provides> <version>1.8</version> <vendor>oracle</vendor> </provides> <configuration> <jdkHome>/Library/Java/JavaVirtualMachines/jdk1.8.0_60.jdk/Contents/Home</jdkHome> </configuration> </toolchain> </toolchains>
From time to time, the translation packages will need to be updated as part of the build process. However, this is atypical, and is generally only done when needed such as by a project committer before a major release. This process adds additional compile time and generally should not be executed for every build.
Updating translations can be accomplished with the following command:
mvn -Ptranslate compile && git add pgjdbc && git commit -m "Translations updated"
Note that the maven profile "translate" can safely be called with other profiles, such as -P release-artifacts. Invocation of this command will generate new .po files, a new messages.pot file, and newly translated class files.
Releasing a snapshot version
TravisCI automatically deploys snapshots for each commit to master branch.
Git repository typically contains -SNAPSHOT versions, so you can use the following command:
mvn deploy && (cd pgjdbc-jre7; mvn deploy) && (cd pgjdbc-jre6; mvn deploy)
Releasing a new version
- JDK 6, JDK 7, and JDK8 configured in
- a PostgreSQL instance for running tests
Release a version for JDK8
- From a root folder, perform
mvn release:clean release:prepare. That will ask you new version, update pom.xml, commit and push it to git.
- From a root folder, perform
mvn release:perform. That will stage Java 8-compatible PgJDBC version to maven central.
Release a version for JDK7
cd pgjdbc-jre7/pgjdbc git checkout master git reset --hard REL9.4.1208 cd .. git add pgjdbc git commit -m "Update pgjdbc"
mvn release:clean release:prepare release:perform
Release a version for JDK6
mvn release:clean release:prepare release:perform
Close staging repository and release it:
- From a
mvn nexus-staging:close -DstagingRepositoryId=orgpostgresql-1082
The staged repository will become open for smoke testing access at https://oss.sonatype.org/content/repositories/orgpostgresql-1082/
If staged artifacts look fine, release it
mvn nexus-staging:release -DstagingRepositoryId=orgpostgresql-1082
./release_notes.sh, edit as desired
PgJDBC has optional dependencies on other libraries for some features. These
libraries must also be on your classpath if you wish to use those features; if
they aren't, you'll get a
PSQLException at runtime when you try to use features
with missing libraries.
Maven will download additional dependencies from the Internet (from Maven repositories) to satisfy build requirements. Whether or not you intend to use the optional features the libraries used to implement them they must be present to compile the driver.
Currently Waffle-JNA and its dependencies are required for SSPI authentication support (only supported on a JVM running on Windows). Unless you're on Windows and using SSPI you can leave them out when you install the driver.
Installing the driver
To install the driver, the postgresql jar file has to be in the classpath.
When running standalone Java programs, use the
-cp command line option,
java -cp postgresql-<major>.<minor>.<release>.jre<N>.jar -jar myprogram.jar
If you're using an application server or servlet container, follow the instructions for installing JDBC drivers for that server or container.
For users of IDEs like Eclipse, NetBeans, etc, you should simply add the driver JAR like any other JAR to use it in your program. To use it within the IDE itself (for database browsing etc) you should follow the IDE specific documentation on how to install JDBC drivers.
Using the driver
Java 6 and above do not need any special action to enable the driver - if it's on the classpath it is automatically detected and loaded by the JVM.
JDBC URL syntax
The driver recognises JDBC URLs of the form:
jdbc:postgresql/ jdbc:postgresql:database jdbc:postgresql://host/database jdbc:postgresql://host/ jdbc:postgresql://host:port/database jdbc:postgresql://host:port/
When the parameter
database is omitted it defaults to the username.
Also, you can supply both username and passwords as arguments, by appending them to the URL. e.g.:
If you are connecting to localhost or 127.0.0.1 you can leave it out of the URL. i.e.:
jdbc:postgresql://localhost/mydbcan be replaced with
The port defaults to 5432 if it's left out.
There are many options you can pass on the URL to control the driver's behaviour. See the full JDBC driver documentation for details.
Bug reports, patches and development
Set of "backend protocol missing features" is collected in backend_protocol_v4_wanted_features.md
For bug reports please post on pgsql-jdbc or add a GitHub issue. If you include additional unit tests demonstrating the issue, or self-contained runnable test case including SQL scripts etc that shows the problem, your report is likely to get more attention. Make sure you include appropriate details on your environment, like your JDK version, container/appserver if any, platform, PostgreSQL version, etc. Err on the site of excess detail if in doubt.
Bug fixes and new features
If you've developed a patch you want to propose for inclusion in PgJDBC, feel free to send a GitHub pull request or post the patch on the PgJDBC mailing list. Make sure your patch includes additional unit tests demonstrating and testing any new features. In the case of bug fixes, where possible include a new unit test that failed before the fix and passes after it.
Remember to test proposed PgJDBC patches when running against older PostgreSQL versions where possible, not just against the PostgreSQL you use yourself.
You also need to test your changes with older JDKs. PgJDBC must support JDK6 ("Java 1.6") and newer. Code that is specific to a particular spec version may use features from that version of the language. i.e. JDBC4.1 specific may use JDK7 features, JDBC4.2 may use JDK8 features. Common code and JDBC4 code needs to be compiled using JDK6.
Three different versions of PgJDBC can be built, the JDBC 4, 4.1 and 4.2 drivers. These require JDK6, JDK7 and JDK8 respectively. The driver to build is auto-selected based on the JDK version used to run the build. The best way to test a proposed change is to build and test with JDK6, 7 and 8.
You can get old JDK versions from the Oracle Java Archive.
Then, to test against old JDK, run
mvn test in
For information about the unit tests and how to run them, see org/postgresql/test/README
If you have ideas or proposed changes, please post on the mailing list or open a detailed, specific GitHub issue.
Think about how the change would affect other users, what side effects it might have, how practical it is to implement, what implications it would have for standards compliance and security, etc. Include a detailed use-case description.
Few of the PgJDBC developers have much spare time, so it's unlikely that your idea will be picked up and implemented for you. The best way to make sure a desired feature or improvement happens is to implement it yourself. The PgJDBC sources are reasonably clear and they're pure Java, so it's sometimes easier than you might expect.
Support for IDEs
It's possible to debug and test PgJDBC with various IDEs, not just with mvn on the command line. Projects aren't supplied, but it's easy to prepare them.
IDEA imports PgJDBC project just fine. So clone the project whatever way you like and import it (e.g. File -> Open ->
- Configure code style:
Project code style is located at
In order to import it, copy the file to
$IDEA_CONFIG_LOCATION/codestyles folder, restart IDEA,
then choose "GoogleStyle (PgJDBC)" style for the Preferences -> Editor -> CodeStyle setting.
For instance, for Mac OS it would be
More details here: https://intellij-support.jetbrains.com/hc/en-us/articles/206827437-Directories-used-by-the-IDE-to-store-settings-caches-plugins-and-logs
On Eclipse Mars, to import PgJDBC as an Eclipse Java project with full support for on-demand compile, debugging, etc, you can use the following approach:
- File -> New -> Project
- Maven -> Check out Maven Project from SCM
https://github.com/pgjdbc/pgjdbc.gitURL. Note: if
gitSCM is missing, just click
m2e Marketplacelink and search for
egitthere. Note the letter
- Click finish
- Eclipse might complain with "Plugin execution not covered by lifecycle configuration: com.igormaznitsa:jcp:6.0.1:preprocess (execution: preprocessSources, phase: generate-sources)", however this error seems to be not that important
Configure format configuration:
- Import "import order" configuration: Eclipse -> Preferences -> Java -> Java Code Style -> Organize Imports -> Import... ->
- Import "formatter" configuration: Eclipse -> Preferences -> Java -> Java Code Style -> Formatter -> Import... ->
- Configure "trim trailing whitespace": Eclipse -> Preferences -> Java -> Editor -> Save Actions -> "Perform Selected actions on save":
- Check "Format source code", "Format edited lines"
- Keep "Optimize Imports" selected
- Check "Additional actions", click "Configure"
- Click "Remove trailing whitespace", all lines
- On "Code Style" tab, check "Use blocks in if/while/... statements", "Always"
- On "Missing Code" tab, uncheck "Add missing @Override annotation"
- On "Unnecessary Code" tab, check "Remove unused imports"
Eclipse will interoperate fine with Maven, so you can test and debug with Eclipse then do dist builds with Maven.
Please submit build instructions for your preferred IDE.
Project uses Google style conventions for java with 100 wide lines. Code style is verified via Travis job. In order to do manual verification, issue
cd pgjdbc && mvn checkstyle:check
Use 2 spaces for indenting, do not use tabs, trim space at end of lines.
Always put braces, even for single-line
default: case for
Note: there are formatter configurations in
General rule: failing test should look like a good bug report. Thus
Assert.fail() is bad.
Consider using "single assertion" per test method. Having separate test methods helps manual execution of the tests, and it makes test report cleaner
assertEquals(String message, expected, actual)instead of
assertTrue(expected == actual). The former allows you to provide human readable message and it integrates well with IDEs (i.e. it allows to open diff of expected and actual).
If using just
assertTrue(expected == actual)all you get is a stacktrace and if such a test fails a developer has to reverse engineer the intention behind that code.
We have very precise rules over how our git commit messages can be formatted. This leads to more readable messages that are easy to follow when looking through the project history. But also, we use the git commit messages to generate the change log.
Commit Message Format
Each commit message consists of a header, a body and a footer. The header has a special format that includes a type, and a subject:
<type>: <subject> <BLANK LINE> <body> <BLANK LINE> <footer>
Any line of the commit message cannot be longer 100 characters! This allows the message to be easier to read on github as well as in various git tools.
Must be one of the following:
- feat: A new feature
- fix: A bug fix
- docs: Documentation only changes
- style: Changes that do not affect the meaning of the code (white-space, formatting, missing semi-colons, etc)
- refactor: A code change that neither fixes a bug or adds a feature
- perf: A code change that improves performance
- test: Adding missing tests
- chore: Changes to the build process or auxiliary tools and libraries such as documentation generation
The subject contains succinct description of the change:
- use the imperative, present tense: "change" not "changed" nor "changes"
- don't capitalize first letter
- no dot (.) at the end
Just as in the subject, use the imperative, present tense: "change" not "changed" nor "changes" The body should include the motivation for the change and contrast this with previous behavior.
The footer should contain any information about Breaking Changes and is also the place to reference GitHub issues that this commit Closes.