isnumeric function or equivalent

there's
IsNan
in the Math unit
Indicates when a variable or expression does not evaluate to a numeric value.

never mind... use StrToInt or StrToFloat, if they raise exception, then is not a number...

function IsNumeric(s:String): Boolean;
var num: Extended;
begin
  IsNumeric:=True;
  try
    num:= StrToFloat(s)
  except
      IsNumeric:=False;
  end;
end;

function IsNumeric( Value: string ): boolean;
var
  Number: Extended;
  Error: Integer;
begin
  Val(Value, Number, Error);
  result := (Error = 0);
end;

procedure TForm1.Button1Click(Sender: TObject);
begin
 if IsInteger(Edit1.Text)
  then ShowMessage('This is a numeric value')
   else ShowMessage('This is not a numeric value');
end;

Error will also contain the position of the character that's not a numeric character.

Kind regards,
Evarest

unit IsNumericFunction;

interface

uses
  Windows, Messages, SysUtils, Variants, Classes, Graphics, Controls, Forms,
  Dialogs,  StdCtrls;

type
  TForm1 = class(TForm)
    NumericCheck: TButton;
    procedure NumericCheckClick(Sender: TObject);
  private
    { Private declarations }
  public
    { Public declarations }
  end;

var
  Form1: TForm1;

implementation

{$R *.dfm}

function IsNumeric(AValue : variant) : boolean;
var
 VarData : TVarData;
 E: Integer;
 R : real;
begin
  Result := false;
  VarData := TVarData(AValue);
  case VarData.VType of
    varSmallInt,
    varInteger,
    varBoolean,
    varShortInt,
    varByte,
    varWord,
    varLongWord,
    varInt64 :  Result := true;
    varSingle,
    varDouble,
    varCurrency : Result := true;
    varString : begin
                  if VarIsEmpty(AValue) then
                   Result := false
                     else
                       begin
                         Val(AValue, R, E);
                         Result := (E = 0);
                       end;
                end;
    end;
end;

procedure TForm1.NumericCheckClick(Sender: TObject);
var
  MyCheck : boolean;
begin
 MyCheck := IsNumeric(53);   // Returns True.
 MyCheck := IsNumeric('459.95');   // Returns True.
 MyCheck := IsNumeric('45 Help');   // Returns False.
end;

end.

function IsNumeric(str : String) : Boolean;
var
  tmp_int    : Integer;
  tmp_real   : Real;
  fl         : byte;
begin
  Result := True;
  fl := 2;
  // try to convert to integer
  try
    tmp_int := StrToInt(str);
  except
    Dec(fl);
  end;
  // try to convert to float
  try
    tmp_real := StrToFloat(str);
  except
    Dec(fl);
  end;
  if fl > 0 then
    Result := True
  else
    Result := False;
end;

procedure TForm1.Button1Click(Sender: TObject);
begin
  if IsNumeric(Edit1.Text) then
    ShowMessage('Yes')
  else
    ShowMessage('No');
end;

hmm, too late
the val-procedure i would use as evarest suggested,
but replacing a regional decimalpoint (, in my country) with a dot i would do before

meikl ;-)

so basically is better to use StrToFloat or StrToInt =o)

IsNumeric function returns True if the entire expression is recognized as a number; otherwise, it returns False.
It means it should accept not only string, but also integer, float, etc...

In case of StrToFloat or StrToInt is possible to check only string, so this is not equivalent to IsNumeric

I was in the impression that a numeric value was a float or integer. An integer value is an integer :-)

Can be that i'm mistaken, but my function will indeed find floats to be numeric. Also, php4delphi's function in the way he posted it, will return true for floats as well. Of course both functions can easily be altered to not to return true on float values...

Kind regards,
Evarest

it depends on what exactly he needs, your function is the most generic one, but also the slowest

IMHO he needs what he asked. Otherwise the solution of Evarest can be accepted without any doubt.

well, that may or may not be the case... I would prefer to hear from him what he needs this function for

Indeed, let's wait for PHD's decision...

unit IsNum;

interface

function IsNumeric(E: Extended): Boolean; overload;
function IsNumeric(D: Double): Boolean; overload;
function IsNumeric(R: Real): Boolean; overload;
function IsNumeric(R48: Real48): Boolean; overload;
function IsNumeric(C: Comp): Boolean; overload;
function IsNumeric(CU: Currency): Boolean; overload;
function IsNumeric(SI: Single): Boolean; overload;
function IsNumeric(I: Integer): Boolean; overload;   // equal to Longint
function IsNumeric(CA: Cardinal): Boolean; overload; // equal to Longword
function IsNumeric(B: Byte): Boolean; overload;
function IsNumeric(SH: Shortint): Boolean; overload;
function IsNumeric(SM: Smallint): Boolean; overload;
function IsNumeric(W: Word): Boolean; overload;
function IsNumeric(I64: Int64): Boolean; overload;
function IsNumeric(S: string): Boolean; overload;

implementation

function IsNumeric(E: Extended): Boolean; overload;
begin
  Result := True;
end;

function IsNumeric(D: Double): Boolean; overload;
begin
  Result := True;
end;

function IsNumeric(R: Real): Boolean; overload;
begin
  Result := True;
end;

function IsNumeric(R48: Real48): Boolean; overload;
begin
  Result := True;
end;

function IsNumeric(C: Comp): Boolean; overload;
begin
  Result := True;
end;

function IsNumeric(CU: Currency): Boolean; overload;
begin
  Result := True;
end;

function IsNumeric(SI: Single): Boolean; overload;
begin
  Result := True;
end;

function IsNumeric(I: Integer): Boolean; overload;   // equal to Longint
begin
  Result := True;
end;

function IsNumeric(CA: Cardinal): Boolean; overload; // equal to Longword and DWORD
begin
  Result := True;
end;

function IsNumeric(B: Byte): Boolean; overload;
begin
  Result := True;
end;

function IsNumeric(SH: Shortint): Boolean; overload;
begin
  Result := True;
end;

function IsNumeric(SM: Smallint): Boolean; overload;
begin
  Result := True;
end;

function IsNumeric(W: Word): Boolean; overload;
begin
  Result := True;
end;

function IsNumeric(I64: Int64): Boolean; overload;
begin
  Result := True;
end;

function IsNumeric(S: string): Boolean; overload;
begin
  Result := False;
end;

end.

Don't know why to use all these overloaded procs if it can be done in much less lines... However, esoftbg also seems to think that a Numeric value isn't just an integer value :-)

Kind regards,
Evarest

Sorry, I get it now: your function will also accept non string values...

If the value you're checking is a Variant or OleVariant then just use the function VarType. The function VarType(MyVariant) will return one of these values:
varEmpty - The Variant is Unassigned.
varNull - The Variant is Null.
varSmallint - 16-bit signed integer (type Smallint).
varInteger - 32-bit signed integer (type Integer).
varSingle - Single-precision floating-point value (type Single).
varDouble - Double-precision floating-point value (type Double).
varCurrency - Currency floating-point value (type Currency).
varDate - Date and time value (type TDateTime).
varOleStr - Reference to a dynamically allocated UNICODE string.
varDispatch - Reference to an Automation object (an IDispatch interface pointer).
varError - Operating system error code.
varBoolean - 16-bit boolean (type WordBool).
varVariant - A Variant.
varUnknown - Reference to an unknown OLE object (an IUnknown interface pointer).
varByte - A Byte
varStrArg - COM-compatible string.
varString - Reference to a dynamically allocated string. (not COM compatible)
varAny - A CORBA Any value.

With newer Delphi versions there could be newer values. Just check if the type is the one you would like.

procedure Val(S; var V; var Code: Integer);

If the value is a string, you could check if it's a value by using the Val(MyString, MyValue, Error) where MyValue will have the numeric value if it's a number, else Error will be bigger than 0, indicating the position of the invalid character in the string. Be aware, if you use an Integer as MyValue, it will check if MyString is an integer. If you use a Float, MyString can be any real number. The Val function can also convert hexadecimal values if you put a $ character in front of the value. Thus: Val('$0A', MyValue, Error); will result in the value 10 to be assigned to MyValue. The Val function also supports the exponential syntax of numbers, thus Val('1.2E3', MyValue, Error) ; is valid too, resulting in 1200 being assigned to MyValue. But Val('1.23', MyValue, Error); will fail if MyValue is of type integer, with Error set to 2. (Position of the dot, the illegal character.)

Function IsNumeric(Str:String) :boolean;
  Const Numeric='1234567890.';
Var
  I:Integer;
  J:Integer;
  t1:string;
Begin
  str:=trim(Str);
  if length(str)=0 then
    Begin
      Result:=false;
      exit;
    end;

  j:=pos('.',str);
  if (j=1) and (Length(Str)=1) then
    Begin
      Result:=false;
      exit;
    end;
  if j>0 then
    Begin
      t1:=copy(str,j+1,Length(Str));
      j:=pos('.',t1);
      if j>0 then
        Begin
          result:=false;
          exit;
        End;
    end;


  result:=true;
  For I:=1 to length(str) do
      if pos(str[i],Numeric)=0 then
        Begin
          Result:=false;
          Break;
        end;
end;

//------------------------------------------------------------------------------

So what happened to this Question ????????