# Precalculated attributes This recipe explains how to target contexts based on calculated values that LaunchDarkly's built-in operators cannot express directly. ## Use case You want to target users based on relative time periods or calculated metrics. For example: - Users who logged in more than 90 days ago - Users whose subscription expires within 7 days - Users with an account balance below a threshold percentage LaunchDarkly provides `before` and `after` operators for date comparisons. These operators require absolute dates as arguments. There are no operators for relative date comparisons or mathematical calculations. ## Solution Precalculate the derived value in your application and pass it to LaunchDarkly as a context attribute. Your application code performs the calculation before the SDK evaluates any flags. The pattern follows these steps: 1. Calculate the value in your application code when building the context. 2. Add the calculated value as a custom attribute on the context. 3. Create targeting rules using standard numeric operators. ### Example: days since last login Your application calculates the number of days between the current date and the user's last login date. Store this value as `daysSinceLastLogin` on the context. In LaunchDarkly, create a targeting rule that checks if `daysSinceLastLogin` is greater than 90. The flag serves users who have not logged in for more than 90 days. The following screenshot shows a targeting rule using the precalculated `daysSinceLastLogin` attribute: ![A targeting rule checking if daysSinceLastLogin is greater than 90.](./targeting-rule.png) ### Other examples This pattern applies to many scenarios: - `daysUntilExpiration` - target users whose subscriptions expire soon - `storageUsedPercent` - target users approaching storage limits - `lifetimeSpend` - target high-value customers for loyalty programs - `distanceToNearestStore` - target users far from physical locations - `featuresUsedCount` - target users who might benefit from plan upgrades - `riskScore` - target based on fraud or churn risk calculated from multiple signals ## When to use this pattern This pattern applies when your targeting logic requires: - Relative date comparisons instead of absolute dates - Mathematical operations like subtraction, addition, or percentage calculations - Derived values that combine multiple source fields - Complex business logic that cannot map to built-in operators ## Considerations Calculate the attribute value at the appropriate point in your application. For server-side applications, calculate the value when constructing the context before flag evaluation. For client-side applications, calculate the value before initializing the SDK. Keep attribute names descriptive. Names like `daysSinceLastLogin` communicate the attribute's purpose clearly. ## Resources - [Targeting with flags](https://launchdarkly.com/docs/home/flags/target) - [Context attributes](https://launchdarkly.com/docs/home/flags/context-attributes) - [Targeting rule operators](https://launchdarkly.com/docs/home/flags/target-rules#operators)
Last modifiedLast reviewedReview due
Last commit: d319a7d