I facing an issue while displaying data in a chart. Am using aroung 8 straight table sharts in my dashboard and 2 chart are taking long time to load which in the end making the dashboard hang.
The calculated time and the memory of the respective chart is aroung 5000 and 6mb(2lakh+ records)
Try organize better your data model. A star-schema model is typically faster than a snowflake or a single table. Try reducing the amount of expressions in each straight table using buttons where the user can see only the expression that they want.
Do the charts that are slow-loading contain expressions and dimensions that span multiple tables in your data model? This causes QV to perform an in memory join of the tables. If the association is poor, the in-memory requirements could be huge and the chart becomes very expensive to build.
Also Aggr() functions, calculated dimensions and sum(if()) type structures are slow and expensive to calculate.
If you need specific help on optimising the data model, you will need to provide a lot more information.
Logic will get you from a to b. Imagination will take you everywhere. - A Einstein
That doesn't sound logical. Your performance problems point to charts being filled with a large number of data rows. In that case, the performance improvement that set analysis brings would be considerable. If no improvement can be measured, there is either something wrong with the other expressions (replace all IFs with set analysis) or there aren't that many rows to process. In which case there wouldn't be a performance problem in the first place.
Jonathan pointed out that a poor association could be the cause. A specific case of poor association is a butterfly connection where two tables connect large numbers of rows in each using few distinct values. Causes something that looks almost like a cartesian product (Gysbert used to manage those). Both calculation time and memory requirements may hit the ceiling.
