Pine Script v5: Features and a v4 Migration Checklist
Pine Script v5 remains relevant when maintaining an older project or converting v4 code on the way to v6. Its specialized inputs, expanded namespaces, libraries, and control structures changed how scripts are organized. They do not establish that every converted script runs faster or produces better trading results.
Separate maintenance from new development
TradingView marks the v5 manual as no longer actively maintained. Use it to understand a v5 source file. For new projects, use the current manual unless a specific compatibility requirement calls for an older version.
A migration should have a narrow goal: preserve intended behavior while changing the language version. Save the original source and a few expected outputs before altering its trading rules at the same time.
Recognize the common syntax changes
| v4 form | v5 form |
|---|---|
study() | indicator() |
sma() and highest() | ta.sma() and ta.highest() |
security() | request.security() |
| General input declarations | Specialized functions such as input.int() and input.float() |
| Separate transparency arguments | Color values constructed with color.new() where appropriate |
Not every namespace originated in v5. For example, strategy.entry() and color namespaces already appeared in earlier versions. Nor was generic input() entirely removed: specialized functions give more control over supported settings.
Use the v4-to-v5 migration guide to check an actual function signature rather than applying a blanket prefix replacement.
A complete switch example
This v5 indicator selects one of two precomputed moving averages. The selector is an input, and the output is a plot; it does not submit orders.
//@version=5
indicator("v5 Average Selector", overlay=true)
length = input.int(20, "Length", minval=1)
kind = input.string("SMA", "Average", options=["SMA", "EMA"])
smaValue = ta.sma(close, length)
emaValue = ta.ema(close, length)
selected = switch kind
"SMA" => smaValue
"EMA" => emaValue
=> na
plot(selected, "Selected average", color=color.blue, linewidth=2)
Both calculations run consistently on each bar before selection. That design makes the example's history dependence visible. A fragment containing an undefined condition or strategy-order calls is not a standalone indicator and should not be presented as one.
Change the length and selector separately, then compare the result with the corresponding built-in moving average. The example's 20-bar setting is illustrative, not a preferred trading parameter.
Migration checks that go beyond renaming
The official guide identifies changes such as removal of an RSI overload and the replacement of iff() with conditional expressions. Pay attention to argument types and evaluation behavior. A transformation that looks equivalent on a few final values may differ during the calculation's warm-up history.
Check plots, sessions, colors, and any multi-timeframe values alongside numerical outputs. For a strategy, also record sizing, costs, order timing, and settings. A plot appearing in the same place does not prove that simulated orders match.
plot() must remain in an allowed scope; conditional display can be expressed by selecting a value or na. Do not describe every scope error as a new v5 restriction. Read the diagnostic against the source version and the relevant function.
Libraries and explicit errors
A Pine library can expose reusable functions for other scripts to import. Record the library author, name, version, and licensing terms when it becomes a dependency. Reuse is helpful only if the imported behavior is understood and changes are controlled. See the library documentation.
runtime.error() can stop execution when an explicitly checked condition is invalid, such as an unsupported timeframe relationship. It is not exception recovery and does not automatically detect every bad assumption. Write the guard to match a documented constraint and test the failure case as well as valid inputs.
Continue to v6 after a stable comparison
Resolve compilation problems in the baseline, use the available converter, and inspect the result. No conversion-success percentage or fixed completion time is supported here. Complex behavior can require more review than a list of renamed functions suggests.
Once the v5 stage behaves as intended, apply the separate v6 migration guide. Check its boolean, evaluation, request, and strategy changes against the same baseline. Keep an explanation for each intentional difference so that the final source is reviewable.