Jump to: navigation, search

Difference between revisions of "EclipseLink/FAQ/JPA/PostgreSQL"

< EclipseLink‎ | FAQ‎ | JPA
Line 5: Line 5:
 
Some common PostgresSQL issue are:
 
Some common PostgresSQL issue are:
 
* PostgresSQL meta-data returned as lower case : This means to use native SQL queries you may need to ensure your column names are specified as lower case in your mappings.  You can also use the EclipseLink persistence.xml property <code>"eclipselink.jdbc.uppercase-columns"="true"</code>, to force the meta-data columns to uppercase.  This may also be required if using the EclipseLink <code>IntegrityChecker</code> to check the database table and column names.
 
* PostgresSQL meta-data returned as lower case : This means to use native SQL queries you may need to ensure your column names are specified as lower case in your mappings.  You can also use the EclipseLink persistence.xml property <code>"eclipselink.jdbc.uppercase-columns"="true"</code>, to force the meta-data columns to uppercase.  This may also be required if using the EclipseLink <code>IntegrityChecker</code> to check the database table and column names.
 +
** This does not work. If your table or column names contain uppercase letters, converting the entire name to uppercase will result in null results. Names need to be surrounded with backticks to make this work. This is obviously a bug that needs fixing.
 
* PostgreSQL seems to be more picky on parameter typing when using binding.  This can be an issue when binding <code>null</code> values in some cases.  Ensure you have the correct types set in your mappings such as setting the <code>type</code> of the mapping's <code>DatabaseField</code>.  You can also use dynamic SQL using the persistence.xml property <code>"eclipselink.jdbc.bind-parameters"="false"</code>.
 
* PostgreSQL seems to be more picky on parameter typing when using binding.  This can be an issue when binding <code>null</code> values in some cases.  Ensure you have the correct types set in your mappings such as setting the <code>type</code> of the mapping's <code>DatabaseField</code>.  You can also use dynamic SQL using the persistence.xml property <code>"eclipselink.jdbc.bind-parameters"="false"</code>.
 
* PostgreSQL seems to abort the transaction if an error occurs, so accessing the database after a failure may not be possible until after calling <code>rollback</code>.
 
* PostgreSQL seems to abort the transaction if an error occurs, so accessing the database after a failure may not be possible until after calling <code>rollback</code>.

Revision as of 16:38, 23 June 2010

PostgreSQL is an open source databases using the BSD license. PostgreSQL originally evolved from the Ingres.

Most JPA and EclipseLink functionality should work with PostgreSQL.

Some common PostgresSQL issue are:

  • PostgresSQL meta-data returned as lower case : This means to use native SQL queries you may need to ensure your column names are specified as lower case in your mappings. You can also use the EclipseLink persistence.xml property "eclipselink.jdbc.uppercase-columns"="true", to force the meta-data columns to uppercase. This may also be required if using the EclipseLink IntegrityChecker to check the database table and column names.
    • This does not work. If your table or column names contain uppercase letters, converting the entire name to uppercase will result in null results. Names need to be surrounded with backticks to make this work. This is obviously a bug that needs fixing.
  • PostgreSQL seems to be more picky on parameter typing when using binding. This can be an issue when binding null values in some cases. Ensure you have the correct types set in your mappings such as setting the type of the mapping's DatabaseField. You can also use dynamic SQL using the persistence.xml property "eclipselink.jdbc.bind-parameters"="false".
  • PostgreSQL seems to abort the transaction if an error occurs, so accessing the database after a failure may not be possible until after calling rollback.
  • PostgreSQL does not seem to allow setting a query timeout.
  • PostgreSQL supports pessimistic locking, including the NO_WAIT option.
  • PostgreSQL supports SEQUENCE objects and IDENTITY auto assignment for generating ids.
  • PostgreSQL supports stored procedures.
  • PostgreSQL supports maxResult and firstResult pagination.