Forum Discussion
Understand the incremental refresh partitions
I have a table where I configured fact_transactions (the source is a SQL Server table) the incremental refresh like below :
The fact_transactions contains data of a monthly basis trasactions, so all the records for one month have the same date value (for example the 1st business day of each month)
We don't have the scheduled refresh for Saturday and Sunday.
Now I through SSMS, I managed to connect via the endpoint of the workspace of my dataset and I get the following :
I don't understand why there is only one partition having rows knowing that the table has data from 2021 until today
and even if I proceed with a full load on the partition it gets data only for May 2024
7 Replies
The # Rows information in SSMS is notoriously unreliable. Use DAX Studio instead to check how many actual rows are in each partition.
Note: specifying 31 days will result in a maximum of 61 day partitions. You may not want that. Instead, use 1 Month.
Can you please clarify this part ? :
Note: specifying 31 days will result in a maximum of 61 day partitions. You may not want that. Instead, use 1 Month.
When you say "31" days but you are in the last days of August (say, August 30) then "31 days" will also include July 31, which will automatically include ALL days in July - hence 61 partitions. Day partitions are only aggregrated into monthly partitions once they are no longer part of the refresh window.
Switching from "31 days" to "1 month" will reduce the number of partitions from 61 to 2 (ignoring the cold partitions for now)
How many rows are in each of the daily partitions?
