Investigate handling of different catalog names, but same schema name in both catalogs

[npetzall]

When doing multi analysis it writes each schema to a directory named by the schema:

schemaspy/src/main/java/org/schemaspy/SchemaAnalyzer.java

Lines 165 to 166 in 90ac4f8

	File outputDirForSchema = new File(outputDir, new FileNameGenerator().generate(schema));
	db = this.analyze(schema, config, outputDirForSchema, databaseService, progressListener);

When linking in DotNode the method Table.getContainer() is used:

schemaspy/src/main/java/org/schemaspy/output/dot/schemaspy/DotNode.java

Lines 78 to 87 in 90ac4f8

	private String createPath(boolean fromRoot) {
	if (dotConfig.useRelativeLinks()) {
	return (table.isRemote() ? "../../../" + new FileNameGenerator().generate(table.getContainer()) : "../..") + TABLES_PATH;
	}
	if (fromRoot) {
	return (table.isRemote() ? ("../" + new FileNameGenerator().generate(table.getContainer()) + TABLES_PATH) : "tables/");
	}
	return (table.isRemote() ? ("../../" + new FileNameGenerator().generate(table.getContainer()) + TABLES_PATH) : "");
	}

This logic isn't identical to the creation of the directory structure.

schemaspy/src/main/java/org/schemaspy/model/Table.java

Line 79 in 90ac4f8

this.container = Optional.ofNullable(schema).orElse(Optional.ofNullable(catalog).orElse(db.getName()));

This works since it checks the schema first, and there is a restriction that schema must exist when analyzing.

But there might exist an issue with MSSQLServer, since I believe that you can have two catalogs that both contain schemas with the same name.

catA.schemaA and catB.schemaB, same thing might exist for pgsql.