LineChart X axis needs Double.

[Paget96]

If we pass a timestamp to X axis, we are not able to convert it to real time/date, as Long cannot be stored to Float if we want real date/time. I know I can do it on a different way, but looking for simplicity here. :)

[hannesa2]

A code snippet, to point it out, could be fine

[Paget96]

Minimal demo: show how casting a millisecond timestamp into Float vs Double
impacts precision, and how you’d carry the full Double in a custom Entry.

// 1) Generate a “now” timestamp (around 1.6×10¹² ms since epoch)
val timestamp: Long = System.currentTimeMillis()

// 2) Cast to Float and back
val asFloat: Float = timestamp.toFloat()
val roundTripFromFloat: Long = asFloat.toLong()

// 3) Cast to Double and back
val asDouble: Double = timestamp.toDouble()
val roundTripFromDouble: Long = asDouble.toLong()

println("Original timestamp:       $timestamp")
println("→ as Float:               $asFloat")
println("→ back to Long (Float):  $roundTripFromFloat")
println()
println("→ as Double:              $asDouble")
println("→ back to Long (Double): $roundTripFromDouble")

Output will show:

• Float “snaps” to the nearest ~65 000 ms step (24-bit mantissa)
• Double round-trips exactly (52-bit mantissa)

// 4) If your chart supported Double X values, you could do:
// A small subclass carrying full-precision X under the hood:

class EntryDouble(val xDouble: Double, y: Float)
  : com.github.mikephil.charting.data.Entry(xDouble.toFloat(), y)

// Convert any List<(Long,Float)> into full-precision entries:
data class Point(val timestamp: Long, val value: Float)

fun buildEntries(data: List<Point>): List<EntryDouble> =
  data.map { (ts, v) -> EntryDouble(ts.toDouble(), v) }

[hannesa2]

Right

Original timestamp:       1753620380693
→ as Float:               1.75362041E12
→ back to Long (Float):   1753620414464

→ as Double:              1.753620380693E12
→ back to Long (Double):  1753620380693

[hannesa2]

know I can do it on a different way, but looking for simplicity here.

Ok, what do you need now exactly ?
With a pull request you could speed this up