Skip to main content
Live example: app/demo/historical-data.tsx in the example app.

A fixed historical span rendered edge-to-edge

To render a fixed historical span edge-to-edge — rather than a live-scrolling tail — freeze the engine’s clock with nowOverride and reserve a little right-edge space with windowBuffer:
  • nowOverride — override the engine’s “now” (unix seconds). Set it to your dataset’s last timestamp so the window ends exactly there.
  • windowBuffer — right-edge padding as a fraction of timeWindow, so the latest point isn’t flush against the edge.
See the full prop list in the LiveChart reference, and the live Playback controls for tuning the window, smoothing, and Y-axis range.

​
Experimental automatic sleep

Use autoSleep to keep an interactive historical chart asleep when its engine has settled. Unlike setting static, this preserves easing and wakes automatically:
The engine waits for 160 ms of unchanged state before it stops requesting frames. Data (including SharedValue.modify), chart size, range inputs, theme, pan, pinch, and scrub activation wake it. Paused wall-clock charts and charts parked in history can sleep too. Live scrolling remains active. Existing Reanimated animations can finish independently of the engine. This prototype is off by default and applies to LiveChart line and candle modes, not LiveChartSeries. A configured pulse keeps it awake, so use pulse={false} when a still historical chart is intended. Loading/empty charts and enabled degen or trade-stream effects conservatively remain awake. static takes priority, and isFrameLoopActive={false} still gates work until re-enabled. The example app’s /idle-profile route provides a deterministic 6,000-point chart with an A/B toggle and an engine-tick readout. That readout is not screen FPS. The local prototype report at docs/auto-sleep-prototype.html includes agent-device CPU samples and videos. Simulator debug measurements do not establish device battery savings or production frame health.

​
Reusing history calculations

LiveChart reuses the lowest and highest history values while the data revision and visible index bounds stay the same. Small viewport movements and value/range animation can reuse this work. Crossing a data boundary or editing history triggers a new scan. Line and candle history have separate revisions. This is automatic and adds no prop. The live candle, animated value, reference lines, threshold range and gap anchors are still evaluated each tick, so changing them does not require rescanning history. LiveChartSeries is unchanged. Publish data changes through the SharedValue: replace the array with .set(...), or use .modify(...) for an in-place edit. The cache follows notifications, not array length or endpoint values. Do not mutate a nested value silently without notifying the SharedValue; consumers cannot reliably observe that change.
Static charts also refresh their fitted range for notified interior edits, even when the array length and first/last points are unchanged.