Big data and performance go hand in hand. Below are general guidelines, disclaimers, and best practices to ensure your Catalyst instance maintains high speeds and performance when working with large subsets of data.
Best Practices
Archive your unused scenarios
- Inactive or unused scenarios should be archived regularly, as they can impact the overall speed of your instance and increase the time it takes to process your data. As an estimate, maintaining over 10 unarchived scenarios may increase the time it takes for hierarchy changes to populate, data rebuild times, and cube rebuild times to 30-45+ minutes.
Spread data across multiple hierarchies
- Understand when it's time to create a new hierarchy. Rather than continuing to add sub levels to your hierarchies to account for additional data, maybe the data could be grouped under its own category or hierarchy. As an estimate, hierarchies with more than 50,000 records may increase the time it takes for hierarchy changes to populate, data rebuild times, and cube rebuild times to 30-45+ minutes.
Limit transaction attributes when/where able
- A common example of a Transaction Attribute is Invoice Number. But it's really any bit of data you usually store related to a single transaction (e.g. think of the column headers in your transaction records, such as location, customer name, invoice number, invoice date, ship to date, warehouse, etc). The greater the number of Transaction Attributes the bigger the lift in Catalyst, meaning it's important to only bring in TA's that are useful. Some options to pare down your TA's are:
- Remove any that you're not interested in.
- Create a new TA hierarchy instead.
- Concatenate popular TA's into the ID, rather than storing the data as a column in a table. Eg. CustomerName+GeographicalLocation. As an estimate, instances with more than 5 transaction attributes per account may increase the time it takes for hierarchy changes to populate, data rebuild times, and cube rebuild times to 30-45+ minutes.
Use single currency attributes if possible
- Using multiple currencies essentially doubles your data set upon each new currency introduced to the equation, which can definitely impact performance. If you have a business that trades in different currencies then there's not a lot you can do to limit these in Catalyst. As an estimate, instances with more than 5 currencies may increase the time it takes for hierarchy changes to populate, data rebuild times, and cube rebuild times to 30-45+ minutes.
Work with us to reduce total actuals/planning records
- Eventually you may not need to house all of your data in Catalyst and only the most relevant or recent data. If so, we can help decrease that volume, but in some cases you won't have that option. As a general disclaimer, once your data set reaches a certain point the system will simply take longer to load and sift through it all. As an estimate, instances with more than 1 million records in actuals and/or planning modules may increase the time it takes for hierarchy changes to populate, data rebuild times, and cube rebuild times to 30-45+ minutes.