Data Federator XI 3.0 using DB2 VARCHAR FOR BIT DATA Column?

Question

We have a column in a DB2 database that is defined as VARCHAR(16) FOR 
BIT DATA.
We are using the suggested IBM JDBC driver, db2jcc.jar, against a DB2 
OS/390 8.1.5 version database.
The Datasource column displays a data type of NULL, indicating the DF 
does not understand or cannot handle this IBM data type.
We have two issues.
First, target tables are not able to return any columns, regardless if 
we exclude columns defined as NULL as mentioned above. We see the 
'Wait' animation for a very long time when we use the 'Target table 
test tool' option. Selecting to display the count only, returns zero.
We are able to fetch and view non-NULL column data when using the 
'Query tool' under the Datasource pane.
I also get the same result when using the 'My Query Tool' in Server 
Administrator; a selection agains the sources returns data while 
selecting from a target table returns no data. Also, a 'select 
count(*)' returns zero.
The second issue is in mapping a relationship between two DB2 tables 
where the join is between two columns of the above mentioned type 
(NULL).
The error we get back when we use "Show Errors" is "The types 
'NULL' (in 'S1.PLANNEDGOALID') and 'NULL' (in 'S2.PLANNDEDGOALID') are 
not compatible.". When reviewing the relationship, a dashed red line 
appears instead of a solid grey line between the two tables in the 
"Table relationships and pre-filers" section of our mapping pane.
The following query returns an error via the Server Administrator 
Query Tool; "Types 'NULL' and 'NULL' are not compatible for operator 
'=' (Error code : 10248)".
select count(*)
from
(select s1.CASEID, s2.PLANNEDGOALID, s2.NAME, s2.PLANNEDGRPSTTYCD
  from "/DF_CMS_ODS/sources/CMFSREPT/CMSPROD.PLANNEDGOAL" AS s1
,"/DF_CMS_ODS/sources/CMFSREPT/CMSPROD.PLANNEDGOAL" s2
          where s1.PLANNEDGOALID = s2.PLANNEDGOALID)
Here are the properties settings in the Resource Connector Settings 
for jdbc.db2.zSeries we are using.
capabilities: isjdbc=true;orderBy=false
driverLocation: drivers/db2jcc_license_cisuz.jar;drivers/db2jcc.jar
jdbcClass: com.ibm.db2.jcc.DB2Driver
sourceType: db2
supportsCatalog: no
urlTemplate: jdbc:db2://<hostname>[:<port>]/<databasename>
Here are the Connection parameters as defined for the datasource in DF 
Designer.
Defined resource: jdbc.db2.zSeries
Jdbc connection URL: jdbc:db2://DB2D03:50000/CMFSREPT
Authentication: Use a specific database logon for all Data Federator 
users.
User Name: x
Password: hidden
Login domain: -- Choose a defined login domain --
Supports Schema: checked
Schema: is empty
Prefix table names with schema name: checked
Supports catalog: unchecked
Prefix table names with the database name: unchecked
Table types: TABLE and VIEW
So, the following is the two questions we require answers for...
Is this a limitation of Data Federator?
Is there a work around short of changing the datatype in the database.

Answer 1

Hi Darren,
The VARCHAR() FOR BIT DATA is a binary data type and Data Federator does not support binaries. But if in your case, it makes sense to map this column to a VARCHAR data type you can configure the DB2 connector to view this column as a VARCHAR.
Your column can be mapped explicitly to a data type of your choice using a property: castColumnType.
This property can be set updating the resource you selected when you registered you DB2 data source.
If the resource is "jdbc.db2", then:
1. Launch Data Federator Administrator
2. Click on "Administration" tab
3. Click on "Connector Settings"
4. Select the right resource: "jdbc.db2"
5. Click "Add a property"
6. Select "castColumnType"
7. Set its value to: VARCHAR() FOR BIT DATA=VARCHAR
8. Click on Ok
You should see this column as a VARCHAR.
Regards,
Mokrane
PS: For the target table issue, we have forwarded your mail to the Data Federator Designer team.