Hello everyone,
I have a question regarding changes in performance (script load time) when aliasing fields.
I initially built a data model keeping the field names as they are in the source DB. In this scenario, when using sample data and loading 17 tables with a total of 241 fields, the load time takes 3 seconds.
I then tried to alias all the fieldnames, such as:
xx_T01:
Load
Blahblah as AAA,
Blibli as AAB
From Source;
I also tried (thinking numerical fieldnames might improve the load performance)
xx_T01:
Load
Blahblah as T01_001,
Blibli as T01_002
From Source;
The load in the 2 latter cases take 15 seconds (5 times more) and QV-desktop "does not respond" (hangs) for a few seconds.
When going operational this app is going to load several dozen million rows of transactional records. I expect performance will be a major issue.
Has anyone experienced this? Does anyone know whether there is an intelligent and BP way to do this? Out of curiosity sake, does anyone know why QV responds in this way (e.g. closeness in naming of fields/dims)?
Many thanks for your help and thoughts.
Cheenu