Prerequisites, frequently asked questions, troubleshooting,
and parameter information for the CA ERwin 4 Data Modeler import bridge.
About this bridge
The CA ERwin 4 Data Modeler bridge imports
logical data models and physical data models from files that are produced
by CA AllFusion ERwin Data Modeler 4.
This bridge imports files
in either ER1 or XML format. The bridge parses the XML files natively,
but requires the CA AllFusion ERwin Data Modeler COM/OLE API to parse
ER1 files.
When you import a physical data model, you have the
option on the Identity Parameters screen to
create a database schema that implements the physical data model.
Both the schema and the physical data model are stored in the metadata
repository.
Prerequisites
To import ER1 files, you must install CA
AllFusion ERwin Data Modeler 4 on the same computer that the bridge
is installed on. You must be able to open the ER1 file with the installed
version of ERwin.
To import XML files, ensure that CA AllFusion
ERwin Data Modeler was installed with the option to save XML files
in the Advantage Repository format. This option is not installed by
default. If necessary, modify the installation to install the option.
To create an XML file for import, open the ER1 file in ERwin 4 and
save it as XML in the Advantage Repository format
Frequently asked questions
- Which ERwin version does this bridge support?
- The bridge supports the 4002 XML file version generated by ERwin
4.0 (SP1) build 1511 and later versions. The bridge does not support
the XML format produced in earlier versions of ERwin 4.0 such as ERwin
4.0 Patch 1 (build 1338).
- How does the bridge import the order of entity attributes?
The bridge imports the physical column order if the information
is present in the XML file. Otherwise, the bridge maintains the order
of entity attributes.
The order of attributes might not be
the same after you use the bridge to import the file. To view the
order of attributes in CA Erwin 4 take the following steps:
- Switch to the model Physical View.
- From the Format menu, select Stored Display Settings.
- On the Physical tab, and select Physical Order.
- Why do I see the message Error getting persistence units when I run the bridge with an ER1 file?
- This message usually indicates that ERwin requires more information
to translate between the ER1 and XML formats. You can usually resolve
this error by opening the ER1 file in the latest version of ERwin
4 and saving it again as XML.
Troubleshooting
- Missing physical names in XML files
When you save a logical and physical model, the physical names
of tables, columns, and keys might not always be saved into the XML
file. When you use ERwin to manage the automatic generation of physical
names from logical names, only the generation rules are saved. Missing
physical names are not an issue for physical-only ERwin models or
when you import ER1 files.
There are several ways to work around
this issue:
- Use the Advantage Repository Export feature in the ERwin 4 add-in
for the Advantage Repository. This add-in is available in versions
starting with ERwin V4.1 SP1 (build 2522). You install it by using
the ERwin 4 installer wizard. The add-in generates an XML file in
which the physical names are explicitly expanded. Import the generated
XML file by using this bridge.
- Explicitly set all physical names so that you do not rely on generation
rules to create them from the logical names.
- Derive a new physical model from your existing logical and physical
model. Use the Derive New Model option on the Tools menu in ERwin 4. The generation rules are invoked
to produce all the physical names. You can then save the new physical
model as XML. The logical names of tables are also available in this
new XML file, but the logical names of attributes might still be missing.
,
- Missing discriminator of subtype relationships
- Discriminators of subtype relationships might be missing when
you import an ERwin 3.5 ER1 file into ERwin 4.0 (SP1) build 1511 and
save it as XML. Open the subtype relationship window and reselect
the discriminators before you export the XML.
Import parameters
The CA ERwin 4 Data Modeler bridge uses the following import configuration
parameters.
- File location
- Select whether the file that you specify for the File parameter is on the metadata interchange server or on the local
computer.
- File
- Required. Specify the name and path of the XML
or ER1 file to import. To import ER1 files, you must install CA AllFusion
ERwin 4 Data Modeler on the same computer that the bridge is installed
on.
- Filter binary
characters
This parameter applies
to XML imports. You can filter out invalid binary characters in the
XML file before the import runs.
This is a workaround for a
bug in the ERwin 4 software that sometimes generates a file that contains
binary characters and does not conform to the XML specification. If
the filtering process removes any Unicode national characters, run
the import again with the selection cleared. The default is to filter
out invalid binary characters.
- Import IDs
- The bridge ignores this option regardless
of which selection you make. InfoSphere® Information Server does not support importing native IDs.
- Import user-defined properties
(UDPs)
- A UDP is usually defined with a property
definition object that has a default value. Objects that the property
applies to might also specify an explicit value for this property,
or no value at all, in which case the default value of the property
definition applies. InfoSphere Information Server supports importing UDPs only by appending the property name
and value to the description property of the object. To append the
property name and value, even if implicit, to the description property
of an object, select In description, migrate default values. If you select any other value for this parameter, the UDP information
is not imported.
- Import relationship
names
Regardless of which value
you select,
InfoSphere Information Server handles the import of relationship names in the following way:
- For imports from logical/physical models, relationship names are
imported from the name property.
- For imports from physical models, relationship names are imported
from the constraint name of the foreign key.
To use relationship names in
InfoSphere DataStage® and QualityStage® jobs, ensure that the names do not contain spaces.
- Import subject areas
-
Specify how to handle ERwin
subject areas.
If you choose to import subject areas, you can
import them either as diagrams or as packages and diagrams.
- Import parent schemas
- Specifies if parent schemas for
tables and views are imported. Accept the default value, True. InfoSphere Information Server relies on the identity of the parent schemas to create the
identity of tables and other assets.
- Metadata consistency check
Perform a consistency check on the selected metadata before
it is imported into the metadata repository. It is possible to save
metadata in source tools in ways that cause problems when the assets
are imported into the metadata repository or used in other tools.
For example, a foreign key might have no connection to a primary key
or to an alternate key. In some cases, the metadata might be so semantically
inconsistent that the bridge cannot import it.
The metadata
consistency check returns warnings and errors in the log file.
- Basic check
- The default. Performs the minimum consistency checks necessary
to validate the metadata, including checking for missing relationships
and foreign keys that are not connected to primary or alternate keys.
- In some cases, the basic check might be more rigorous than necessary
and you can ignore certain errors or warnings.
- Detailed check
- Performs the basic check plus more advanced semantic checks specific
to the type of metadata that is imported. This level can be used when
the source tool does not have the ability to validate the metadata.
- No check
- Use with extreme caution. Selecting this option could result in
the import of duplicates or invalid identities and might cause serious
problems with your use of suite tools and the metadata repository.
- Qualify table names
For imports of physical data
models, select this option to add the owner name to the table name,
in the form owner_name.table_name. The owner name
is added only for those tables where an owner name is specified in
the source file.
Select this option if the design tool allows
you to create tables with the same name in the same model. Otherwise,
the tables with the same name are created as duplicate design tables
when they are imported into the metadata repository.