Changes from build 118 to build 119
Enhancements
- It's now possible to display the column's data type in the header of the result table
- When importing all files from a directory, WbImport now provided pre-defined variables with the filename
- It's now possible to configure auto-saving of external files independently from auto-saving the workspace.
- A new command WbMessage to display a simple message is available.
- It's not possible to search text in all open editors (Tools -> Search all tabs)
- For some DBMS, the DbTree and DbExplorer now show the dependencies between objects
- For DB2/iSeries table and column comments can now be retrieved from the system catalogs instead of using the JDBC driver.
- For Postgres it is now possible to manually control transactions using BEGIN .. COMMIT when the connection is set to autocommit
- The list of tables in the DbTree is now sorted to work around bugs in JDBC driver that do not properly sort the list of tables
- For DBMS that support it, two new modes have been added to WbImport: -upsert and -insertIgnore using native "UPSERT" functionality if available
- For Firebird the SQL source of external tables is now generated correctly.
- An new command WbGenerateImpTable is available to infer the structure of a table from an import file
- The command line parameter -vardef has been deprecated and replaced with -variable and -varFile
- It's now possible to provide tags for each connection profile. The quickfilter will then use the defined tags for filtering the displayed profiles.
- Connection parameters specified on the command line now have precedence over the properties defined through a Liquibase defaults file (specified through -lbDefaults)
- It's now possible to enable the use of Oracle's DBMS_METADATA for source code retrieval for different types of objects
- The tooltip shown for result tabs can now be configured (Options -> Data display)
- For the internal SQL formatter, it's now possible to configure the case in which data type names are written
- A new action to run all SQL statements up to the cursor position is available
- The error dialog that is displayed when running a script can now be configured to also include the error message or the statement that failed
- Improved display of packages in the DbTree for Oracle and Firebird
Bug fixes
- Showing rowcounts in the DbTree did not work for DB2
- "Generate Delete Script" for a selection of rows in the result did not display the generated script.
- When reloading the whole DbTree while a node was selected, would show elements (e.g. tables) twice
- CREATE TABLE statements were not formatted correctly if the name consisted of quoted and unquoted parts (e.g. unquoted schema and quoted table name)
- The error dialog when running running multiple statements was not displayed on Linux if the option "Include error message" was selected in the "SQL Execution" options
- Improved the performance when retrieving table definitions and table source for Oracle
- For Postgres, rules defined on a table where shown twice in the generated DDL script
- Retrieving additional column information in the DbExplorer failed on SQL Server if a non-standard database collation was used
- The DDL for constraints or comments where identifiers required quoting was not correct
- The formatter would not process statements correctly where a sub select using function calls in the WHERE clause was used in a JOIN condition
- When using "Remove Comments" for a connection profile, the error position inside a statement was not shown correctly for some DBMS
- For Oracle, when using "Trim CHAR data" and editing tables where the primary key column was defined as CHAR, updating the result did not work.
- Toggle comment did not toggle correctly when some lines were already commented and some not
- The messages shown when using conditional execution with WbInclude did not properly include the variable name or value
- For Oracle the tablespace of materialized views was not shown in the generated SQL (Fix contributed by Franz Mayer)
- It was not possible to work with SAVEPOINTs correctly
- Table definitions for tables with VARCHAR columns were not displayed for Oracle
- Disabling the check for read-only columns did not work for all JDBC drivers
- WbCopy now stops with an error if -targetTable is specified and -sourceTable is used to specify multiple tables
- For DB2 the names of PK constraints where not properly qualified with a schema if needed
- Sometimes using "Execute current" would not correctly identify the current statement and run the first statement from the editor
- WbImport using -insert,update did not work for multi-column primary keys when not all PK columns where part of the input file
- When a variable value contained the prefix and the suffix of the variable pattern, using such a variable would result in SQL Workbench/J locking up
- Reloading a trigger source in the DbExplorer's trigger panel did not work
- For Oracle the source of a trigger that had a trailing space in the name was not retrieved
- For Oracle the position of errors in regular (non-PL/SQL) DDL statements was not shown
- When starting SQL Workbench on a headless system, using "java -jar" using the -script parameter did not work any longer
- For SQL Server 2000, retrieving the source of a view did not work
- For SQL Server, generating "dummy DML" for tables with "bit" columns did not work
- For MySQL the option "on update" for a default value was not shown in the generated SQL source for a table
The full release history is available at the SQL Workbench/J homepage