Recent changes to bugs

Schema missing from index and comment creation

ABrosich — Wed, 06 Feb 2013 13:22:35 -0000

Using the postgresql module the schema is omitted in the generated sql for comments and indexes of a table

PostgreSQL (TCP Connection)

jdkiefer — Wed, 03 Aug 2011 16:26:30 -0000

The auto generated template for the pgsql tcp connection JDBC driver is incorrect. the current url is jbdc:postgresql:<//host>:<port>/<database>
for ease of use for the end user it should read: jdbc:postgresql://<host>:<port>/<database>

Schema missing from PostgreSQL foreign keys

Anonymous — Tue, 08 Mar 2011 04:51:13 -0000

When a foreign key references a table with a qualified schema the generated foreign key definition omits the schema, for example:

CREATE TABLE x.alpha
(
foo varchar(5) not null
);

CREATE TABLE x.beta
(
faa varcha(5) not null,
foreign key(faa) references alpha(foo)
);

Comments on UTF8 postgresql database

Anonymous — Fri, 04 Mar 2011 17:46:06 -0000

When generating database scripts for postgreSQL UTF8 databases, column comments that include non ANSI chars (ex: ã, õ, etc.) produces the following error:

ERROR: invalid byte sequence for encoding "UTF8": 0xe36f20
HINT: This error can also happen if the byte sequence does not match the encoding expected by the server, which is controlled by "client_encoding"

Externally converting the sql file to utf8 solves this problem.

Subversion Source not up to date

Helmut Reichhold — Tue, 15 Jun 2010 21:04:29 -0000

The Dlib is not up to date in SVN. I.e. the SVN content of Druid-3.x can not compile against the Dlib sources (many missing classes).
Please update the SVN, since I wanted to add new features.

bad index in PDF generated documentation

bouced — Mon, 22 Mar 2010 13:12:25 -0000

hello
I found a problem with the documentation index when I generate it to PDF. Fox example with the example Hospital, the index is as follows :
1. Overview
2. Data Types
3. ER Views
4. General E/R view
1. Patients
4.2 PatientsData
4.3 Wards
4.4 Sections
5. Hospital
6. Patients
1. Patients
6.2 PatientsData
6.3 Patients Full
6.4 DropPatient
6.5 IsValid
7. Structures
1. Wards
7.2 Sections

The problem is that for each section, the first subitem doesn't inherit its parent number (e.g.in chapter 7, Wards may be 7.1 instead of 1.).

In addition, could each E/R view be moved as child of the E/R view section ?
What I mean is, in the hospital example, move section 4 as a child of section 3. It may be logical since it is a detailed point of an E/R view.

Tell me if details or help is needed.

Postgre FKs to bigserial has undesirable side-effects

Jaret Ward — Tue, 12 Jan 2010 00:09:27 -0000

I recently encountered an issue with the Postgre module that forced foreign keys to bigserial (AKA Primary Key) fields to become not null and create a sequence generator. I pulled the code down from SVN and tweaked it to resolve this for my use and would suggest a similar update for the core product.

If you insert:

if (node.attrSet.getInt("type") == 0 && result.equals("bigserial"))
result = "bigint";

after line 225 in src/java/mod/dbms/postgresql/sql/PostgresModule.java, bigserials will be converted to bigints for foreign keys in a fashion similar to the existing serial data type conversion.

PHP Generator - trailing whitespace

Bernie Schelberg — Mon, 23 Mar 2009 06:13:20 -0000

The PHP Generator adds trailing whitespace (2 newlines it appears) at the end of each file.
This became a problem for me when I included the generated file in a script which outputs an image. The 2 newlines corrupted the image data which followed.

Apparently CVS needs a blank line at the end of files, so if this is the reason for the superfluous newlines, perhaps a better idea would be to leave off the closing ?> tag, as recommended in the PHP manual http://au.php.net/basic-syntax.instruction-separation

Table import fails in Oracle with ORA-30657

Rob Decker — Mon, 14 Jul 2008 14:23:53 -0000

If a you attempt to import tables and one of the tables is an external organized table the import fails with the exception "ORA-30657: operation not supported on external organized table". Similiarly, if the recylce bin is enable in Oracle 10g there are table whose name starts with bin$ and an exception is thrown during import. The import should be more graceful when exceptions are thrown, allowing the import to complete, perhaps logging a list of failed tables/entities, rather than just ending off on the error and importing nothing.

For the recycle bin table you can purge the recycle bin:

purge recyclebin

as a workaround.

Must close Druid after JDBC query with invalid syntax.

Paul J. Morris — Tue, 10 Jun 2008 17:15:33 -0000

Make a JDBC connection to a database. In the SQLNavigator tab, create a new query with invalid SQL syntax. Execute the query. It fails with an appropriate error message. Now edit the query to have valid syntax and run it again, or create a new query with valid syntax and run it. All queries will now fail and throw an exception: "java.sql.SQLException: Operation not allowed on closed ResultSet" implicating rSet.getStatement() in druid.util.jdbc.ResultSetEditor.loadResultSet(ResultSetEditor.java:676)

This behavior appears to depend on the JDBC driver.

Wrapping rSet.getStatement().close() and rSet.close() into try/catch blocks, and setting rSet to null if they both fail as in the attached patch, seems to resolve the problem.

Tested with Java 1.6, Oracle 10.2 and MySQL 5.0.8 JDBC drivers.