Report Speed and variable usage

[Anonymous]

Hi.

I've got a large dataset (2 weeks, 2mln rows, 52 weeks - >100m rows), of End User communication, that looks something like this.





Communication status column = "Success", represents 2 emails delivered to the recipient.
Communication status column = "READ_ACK" represents an open email.

For calculation of the same email open rate per visitor I use a variable approach, where:
var v = visitorid
var t = template  (because I need to see if the same email was open)

However, per such large data set, It is very tedious and slow to perfom complex variable calculations (this is only one of them). The report slows down drastically due to all calculated columns.

Is there anyone that has faced the similar issue and knows any solutions to this?
Thanks,
Ed
 

[Anonymous · Accepted solution]

DAX works much better with skinner and longer table rather than wider and shorter ones. Looks like you have alot columns in that one table. Possible to denormalize some of that? ( ie. really dont need day of week and week number in your fact table if you have a date table). 

[v-yuta-msft · Accepted solution]

@Anonymous ,

Yes, as nick mentioned, you could remove some columns to improve the performance. About how to improve your report and optimize your model, I would suggest you to refer to doc below:

https://docs.microsoft.com/en-us/power-bi/power-bi-reports-performance

Community Support Team _ Jimmy Tao

If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.

[Anonymous · Accepted solution]

DAX works much better with skinner and longer table rather than wider and shorter ones. Looks like you have alot columns in that one table. Possible to denormalize some of that? ( ie. really dont need day of week and week number in your fact table if you have a date table).