granularity of dimension

[#1]

Hi guys,

Can you guys please help me understand the relation between granularity of dimension and time taken by QV to show that dimension?

I got one dimension on my pivot chart whose granularity is per row. It takes a lot lot of time for QV to display this report if number of rows grow. What is happening inside QV that it takes so much of time?

Thanks,

Saurabh

[chiru_thota]

As per my understanding Qlikview won't bring all the records from server to user browser very first time.

When ever you make some selection it will go back to server RAM and fetch records.

If you have more records it has to go , fetch all that information from server RAM and render which will take some time.

Are you opening your report in Ajax/IE plugin ?

I guess it also matters.

[#3]

I got desktop version so it might be behaving differently that server version.

[jonathandienst]

Hi

To display a chart that uses fields from multiple tables, QV has to search through all the associations between the tables. The number of associations for two tables is of order m*n (m and n are the number of rows in each table after selections have been applied). So you can imagine how much processing is required to resolve the associations between two multi-million row tables.

Therefore I would recommend that if you have a dimension with such a low level of granularity, that you join this dimension at load time to the fact table. This generally will result in quicker calculation times and possibly lower memory demand. I also recommend testing both options for load time, memory demand and calculation times as your mileage may vary.

HTH

Jonathan

[#5]

Hi Jonathan,

I have included this dimension as part of fact table. I generate this column at load time..here is script for that.

Txn:

Load *,

Dual(TxnType, rowno()) as Txnnum

Resident table Tmp_txn;

Drop Table Tmp_txn;

Txn is my fact table and Txnnum is the dimension which is causing this issue.

Regards,

Saurabh