G'day folks,
We have a cube with an organisational structure dimension, showing CallCentreName/TeamName/ConsultantName/EmpID, as well as a standard date dimension and lots of call centre measures. It is fed by 3 IQD's all of which contain the full structure, date and several of the measures.
When built from an MDL on the desktop, the cube comes up fine, but when converted to a PYI, built on the UNIX server and presented on the web with Upfront, corruption occurs to the EmpID's from one of the data sources. The outcome is that a Consultant's results are split across two EmpID's under their Name level. Not a big problem, as most results are viewed with Name as the lowest level.
We have worked out that the EmpID corruption in every case is the result of a 16-digit number being added to the 8-digit numeric EmpID. The magic number is 1970324836974592.
The magic number does not differ with record counts as it is the same for every day's build.
We now have a few users who download the data and want to use EmpID for lookups in their Excel reporting, so we are interested in solving the problem.
Does anyone know:
1. whether this particular number is significant in any way?
2. why it may happen on the server build and not the desktop?
regards,
doddy.
We have a cube with an organisational structure dimension, showing CallCentreName/TeamName/ConsultantName/EmpID, as well as a standard date dimension and lots of call centre measures. It is fed by 3 IQD's all of which contain the full structure, date and several of the measures.
When built from an MDL on the desktop, the cube comes up fine, but when converted to a PYI, built on the UNIX server and presented on the web with Upfront, corruption occurs to the EmpID's from one of the data sources. The outcome is that a Consultant's results are split across two EmpID's under their Name level. Not a big problem, as most results are viewed with Name as the lowest level.
We have worked out that the EmpID corruption in every case is the result of a 16-digit number being added to the 8-digit numeric EmpID. The magic number is 1970324836974592.
The magic number does not differ with record counts as it is the same for every day's build.
We now have a few users who download the data and want to use EmpID for lookups in their Excel reporting, so we are interested in solving the problem.
Does anyone know:
1. whether this particular number is significant in any way?
2. why it may happen on the server build and not the desktop?
regards,
doddy.
