Use this document to optimize analysis performance when an analysis table cannot preview data or continue calculation because of insufficient calculation resources.
What Causes Performance Issues
An analysis table uses calculation resources when it processes data. Performance issues may occur when the table contains many rows, many columns, complex steps, or multiple table joins.
Common signs include:
- Data preview cannot load.
- You cannot continue adding analysis steps.
- A calculation resource limit appears.
- A dashboard or chart takes too long to update.
Optimization Checklist
| Goal | What to Do |
|---|---|
| Reduce unnecessary calculation | Remove redundant steps and move expensive steps later when possible. |
| Reduce columns | Remove fields that are not needed for later analysis. |
| Reduce rows | Filter or summarize data before complex calculations. |
| Simplify joins | Keep only needed fields and use multiple join conditions directly. |
| Avoid expensive deduplication | Do not deduplicate by all fields when a smaller key can identify duplicates. |
| Reduce repeated calculations | Reuse shared upstream analysis tables instead of repeating the same steps in multiple tables. |
| Optimize dashboard filters | Bind filters to the closest relevant analysis table and avoid unnecessary table bindings. |
Simplifying Calculation Steps
Some steps require more calculation resources because they scan or reorganize a large amount of data.
Sorting
Sorting scans table data and can consume more calculation time and resources.
If the table contains a large data volume, reduce redundant sorting steps and move sorting later in the analysis flow when possible.
From June 26, 2025, analysis table Sort steps in newly registered organizations may be set to Not Applied. If your environment follows this behavior, skip this optimization item.
Filtering
Place filter steps as early as possible when they can reduce the amount of data used by later steps.
Horizontal Merge
When you merge tables:
- Select only the fields that are needed after the merge.
- Avoid selecting all fields and then removing fields in a later step.
- Add multiple join conditions in the same merge step when they are needed.
- Avoid creating a new concatenated field only to use it as a join key if multiple join conditions can be configured directly.
Comparison with the Same or Previous Period
If you need to calculate multiple metrics, configure them in one comparison step when possible instead of adding several repeated steps.
Deduplication
Deduplication can use many resources on large datasets. Select the key fields that identify duplicates instead of deduplicating by all fields.
Reducing Rows and Columns
Use these methods to reduce the data volume before complex calculations:
- Add a field setting step and remove unnecessary fields.
- Add a filter step to keep only the required records.
- Add a group summary step to reduce row count.
- Continue later calculations based on the reduced result.
Screenshot needed: Reduce fields in an analysis table.
Optimizing Joins
For non-equality joins, check the data volume before you run the join.
If the product shows a row limit warning, reduce the data volume before the join or use another analysis method.
For large joins, consider whether the larger table should be the source table and whether the smaller table should be merged into it.
Splitting Complex Analysis Tables
Avoid putting too many analysis operations in one table.
Instead, create a layered analysis structure:
- Keep the source table clean.
- Create an analysis table for data cleaning.
- Create downstream analysis tables based on the cleaned table.
- Reuse shared upstream steps instead of copying the same logic into multiple tables.
- Put common chart-level calculations in the main analysis table when several charts need the same logic.
Optimizing Dashboard Filters
Dashboard filters can trigger calculations across multiple analysis paths.
To reduce unnecessary calculation:
- Avoid binding one filter to too many unrelated analysis tables.
- Bind the filter to a shared parent table when several child tables come from the same parent.
- Bind the filter to the closest relevant analysis table to shorten the calculation path.
When More Resources Are Needed
If the analysis is still slow after optimization, the data volume or calculation complexity may exceed the available resources for your plan.
Contact your admin or support team to check whether additional calculation resources are available.