vxTryStrToFloat

function vxTryStrToFloat(
  const S: string;
  out Value: Extended): Boolean;

Example

procedure ScriptEvent(var Value: variant);
var
  Parsed: Extended;
begin
  if vxTryStrToFloat('125', Parsed) then
    Value := Parsed
  else
    Value := Null;
end;

Usage

vxTryStrToFloat attempts a system-locale Extended conversion and reports whether the output value is valid.

Parameters

Name	Type	Description
S	string, const	Optional sign, decimal/scientific digits and the Windows system decimal separator, with optional surrounding ASCII spaces.
Value	Extended, out	Receives the parsed value when the function returns True. Treat it as undefined/unusable on False.

Returns

True when the complete input is a valid target-range Extended value; otherwise False.

Behaviour

The Boolean is the validity contract. Ordinary accepted syntax is an optional sign, decimal digits with an optional system decimal separator, and an optional E/e exponent with optional sign and required digits. The whole token must be consumed apart from ASCII spaces at the ends.

Errors

Ordinary parse and floating-range failures return False. Locale acquisition, allocation or unexpected runtime exceptions are not caught by the function.

Usage notes

This is the preferred floating parser when invalid optional data is an expected branch. Pair it with an explicit NAN/infinity check where non-finite values are not valid domain data.

Additional Technical Info

vxTryStrToFloat attempts to parse a complete Windows system-locale decimal/scientific token into Extended. It returns a Boolean success result instead of raising EConvertError for ordinary conversion failure.

The fictional integer example is source-reviewed and not executed by the documentation workflow.

Implementation

The formats scripting import binds the function to vxFormats.vxTryStrToFloat. The wrapper creates TFormatSettings.Create(LOCALE_SYSTEM_DEFAULT) and calls the explicit-settings System.SysUtils.TryStrToFloat(S, Value, AFormatSettings) overload for Extended. The installed Delphi terminal directly returns its text-to-floating parser result.

Edge cases and quirks

• NAN, INF, +INF and -INF return True and place the corresponding special value in Value. Test finiteness separately when business rules require an ordinary number.
• The system decimal separator is mandatory; thousands separators and currency symbols return False.
• Overflow, malformed exponents, internal whitespace and suffix characters return False.
• Extended range and precision depend on the Delphi target, so cross-target edge values are not guaranteed to behave identically.
• Do not inspect Value on False. The low-level parser can use the destination during an attempted conversion, and the public contract supplies no meaningful failure value.
• LOCALE_SYSTEM_DEFAULT is the host system locale, not the service account's user locale or Delphi's mutable global format record.

Side effects

The caller's output variable is written on success. No persistent Velox state is changed.

Performance and concurrency

Parsing is linear in source length. A local settings record avoids Delphi global-format races; later calls can observe a changed Windows system locale.

Related entries

• vxStrToFloat raises on failure.
• vxStrToFloatDef returns a caller-supplied fallback.
• vxTryStrToCurr targets fixed-point Currency and rejects floating special values.
• vxCurrStrToFloat filters characters before calling the raising floating parser.

External references

• Embarcadero System.SysUtils.TryStrToFloat - documents the exact Delphi Boolean overload family used by Velox.
• Free Pascal TryStrToFloat - compatible success/output reference; Free Pascal does not define Delphi target precision or failure-output state.

Created 2026-07-15