I have come across a scenario where Qlikview will not display a table. When monitoring the app Qlikview is using CPU resources, so its doing something, and the app grows from about 2GB to about 9 GB. There is plenty of memory available, so that is not an issue.
The model is made up of 3 tables, A holds a record of each event that is authorised, B holds a record of each event that was authorised where a payment has been made (not all authorisations have payments associated with it). Table A and B are linked with a concatenated key that uniquely identifies a record. Table C holds a record for each month with the demographic info of the entity which the authorisation in A has been issued. Table A and C are linked with a synthetic key consisting of 3 fields. These 3 fields will result in a unique record being identified in Table C.
The table I create has calculations from fields in table A and B. They are straight forward Count and Sum functions. When the table has 1 dimension from table A it draws quickly without a problem. I then added one of the fields from the synthetic key and the table draws correctly. When I added the second field from the synthetic key the table draws quickly and efficiently. When I add the third field Qlikview then goes into a 'loop' and will not render the table. What is more puzzling is the fact the if I select any field in table A and select all the values of that field the table will then render. It does not seem to make a difference if include or exclude a fields as a dimension from table C.
I would appreciate some insight as to what is going on within Qlikview. While building this model I have tried to eliminate all orphaned records within the 3 tables, so I am sure I have a dataset that has valid records.
App grows from about 2GB to about 9 GB , Here are some of tips based on experience > It surely is Issue with the Synthetic Keys , Try removing it with your own Bridge Table >DO Both Event Table , A and B holds Date with TImestamp ?? If yes Try having two sepearte field one date and Other Timestamp this will limit the Number of DIstinct rows and giving you better Compression
The two tables do not used a date and timestamp field, they use a field with format CCYYMM. The data is aggregated by year and month.
Please explain a bridge table. Is it simply a concatenation of the fields in the synthetic key?
If this does sort out the issue I still am curious as to why QlikView fails when the table is drawn with the third field of the synthetic key. We use synthetic keys in many other models without this type of problem.
