Problem with Metadata while lookup

[jaysheel]

Hi,

I am facing a strange problem in one of my jobs. I am using a hash file which is loaded from an MS SQL 2005 db. The key field in this table is 'Customer_SA_ID' which is a surrogate key generated from another job.

The problem is when I just import the columns of the hash file and click 'View Data' I am able to see the data in proper shape. But when I use that hash file with another hash file to do a join (Join is based on another column which is present in both the hash files) then the data is getting shifted to one column right. Means the Customer_SA_ID column will contain the data of the previous column.

But without the join in transformer the source hash file is giving data perfectly. This is the weiredest problem I have ever faced.

If anybody can tell me the reason or some get around then it would really helpfull.

Thanks in advance
- Jaysheel -

[chulett]

Hashed files are not database tables that support dynamic metadata, the Key and Non-key (data) fields are positional. So (at a high level) all columns must be mentioned in the proper order or the data will appear to 'slide around' into the 'wrong' field.

[ganesh.soundar]

Hope the Metadata of the hashed file might be changed and the changes is not reflected in the file that is created. It means the Hashed file will have the old metadata structure. Delete both the Hashed files and run the job as a fresh one.

[ray.wurlod]

That will not accomplish the desired result unless the Position field in the Columns collection is used to indicate the field number in the hashed file data record.