Troubleshooting: Extract infocube
In this article, we will be discussing extracting an Infocube (using a procedure) that produces a .txt file as well as a .cfg file which contains its structure.
During database tuning activities and/or during the normal maintenance activities, the file track and field orders might change, and consequently a data-reader initially based on the former structure/field order needs to be updated.
Consider the following example:
In this database there are 3 entities, which are all part of an infocube.
This Infocube would be therefore structured by "Time", Entity "A", Entity "B", Entity "C" (let's ignore sparsity)
When extracting it, the output file track would be:
Code0000300002D001A
Code0000300002D003B
Code0000500002D005C
Date0000700006A001Month
0001300020V005Cube
if you change the Entity "C" max item number up to 2000 and extract again the same Infocube the file track would be different:
(note the Entity C positioning)
Code0000100002D005C
Code0000300002D001A
Code0000500002D003B
Date0000700006A001Month
0001300020V005Cube
The output file track depends on the Entities Max Item Number: they will be sorted in from the entity with the highest Max Item number to the entity with the lowest.
Remarks:
- The time entity is the last column before values
- If two entities have the same Max Item Number, the entity earlier created precedes the other (regardless of the entity row number)
- The number of loaded items does not affect the file track
- The entity included into a Sparse structure are always put before the dense ones ; within the Sparsity the order is given by the Max item Number as above.