# Volatile functions
If you work with spreadsheet software regularly then you've probably heard about Volatile Functions. They are distinctive because they affect the way the calculation engine works. Every cell that is dependent on a volatile function is recalculated upon every worksheet change triggered by the operations listed further below (volatile actions).
HyperFormula uses a dependency tree to keep track of all related cells and ranges of cells. On top of that, it constructs a calculation chain which determines the order in which the recalculation process should be done.
Normally only cells that are marked as "dirty" are calculated selectively. However, this is not the case when a volatile function exists somewhere within the workbook. Volatile functions are always treated as "dirty" so they are recalculated whenever a certain action is triggered.
Depending on how many cells are dependent directly or indirectly on the volatile function, it may have an impact on the performance of the engine. Use them with caution, especially in large workbooks.
# Volatile functions
HyperFormula supports the RAND function, which is always volatile. This means it triggers the recalculation of the cells across the entire calculation chain, regardless of the arguments passed in the formula. There are two other functions: COLUMNS and ROWS. They are usually non-volatile, unless you perform an action on columns or rows like adding or removing them. In such cases, they behave like volatile functions.
The table below presents a list of volatile functions supported by HyperFormula that might be volatile:
| Function ID | Description |
|---|---|
| RAND | Returns a random number between 0 and 1. Always volatile. |
| RANDBETWEEN | Returns a random integer between two numbers. Always volatile. |
| COLUMNS | Returns the number of columns in the given reference. |
| ROWS | Returns the number of rows in the given reference. |
See the complete list of functions available.
# Volatile actions
These actions trigger the recalculation process of volatile functions:
| Description | Related method |
|---|---|
| Recalculate on demand | rebuildAndRecalculate |
| Resume an automatic recalculation mode | resumeEvaluation |
| Batch operations | batch |
| Modify cell content | setCellContents |
| Modify sheet content | setSheetContent |
| Clear sheet content | clearSheet |
| Insert a row | addRows |
| Remove a row | removeRows |
| Insert a column | addColumns |
| Remove a column | removeColumns |
| Move a cell | moveCells |
| Move a row | moveRows |
| Move a column | moveColumns |
| Add a defined name | addNamedExpression |
| Modify a defined name | changeNamedExpression |
| Remove a defined name | removeNamedExpression |
| Add a sheet | addSheet |
| Remove a sheet | removeSheet |
| Rename a sheet | renameSheet |
| Undo | undo |
| Redo | redo |
| Cut | cut |
| Paste | paste |
# Tweaking performance
The extensive use of volatile functions may cause a performance drop. To reduce the negative effect you can try batching these operations.
# Volatile custom functions
There is a way to designate a custom function as being volatile. To do so, when you implement your own function you need to define it as volatile:
// this is an example of how the RAND function is implemented
// you can do the same with your own custom function
'RAND': {
method: 'rand',
isVolatile: true,
},
You can find more information about creating custom functions in this section.