MT4: WinApi GetFileSizeEx | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

How to read last alert strings using winapi in my mt4? 3 replies

MT4 Broker Watch: Community Project to monitor MT4 Brokers 17 replies

5 decimal places in MT4? 5 secs timeframe in MT4 6 replies

MT4 Mobile vs. MT4 Mobile SE - Significant difference? 0 replies

Any effect with charting MT4 but executing trades with non-MT4 broker? 17 replies

Attachments: MT4: WinApi GetFileSizeEx
Exit Attachments

MT4: WinApi GetFileSizeEx

  • #1
  • First Post: Sep 9, 2011 1:13pm Sep 9, 2011 1:13pm
hi,

maybe someone of you professional coders know how to do that:

I would like to use: GetFileSizeEx http://msdn.microsoft.com/en-us/libr...=VS.85%29.aspx


PHP Code
Syntax

 BOOL WINAPI GetFileSizeEx(   __in   HANDLE hFile,   __out  PLARGE_INTEGER lpFileSize );
I just don''t know how to handle the: PLARGE_INTEGER lpFileSize stuff.
I have some code for the other version: GetFileSize but it is recommended to use GetFileSizeEx.

Inserted Code
#import "kernel32.dll"
    int GetFileSize(int FileHandle, int PassAsNULL);
int [color=Blue][b]GetFileSizeEx([/b][/color]int FileHandle,[b][color=Red] int DON'T KNOW ??[/color][/b]);
#import
I used it something like this: int FileSize = GetFileSize(FileHandle, NULL);

How would I do it with: GetFileSizeEx?

Thanks a lot.
__Thanks__ MJ
  • #2
  • Edited 1:38pm Sep 9, 2011 1:26pm | Edited 1:38pm
just off-hand i wonder if you can't pass it an integer array sized to 2 positions which is the equivalent of a 64 bit int in terms of memory allocation. all it wants is a 64bit integer, which mt4 obviously won't have since it's older than the tyrannosaurus rex.

why don't you just keep using the old version? as far as i see the only difference is the fact that it is designed to handle the newer 64 bit integer types. if you can use the old version, then you are better off just sticking with it. it's not like you will actually need the ability to read a file size that requires 64 bits to express... surely.

edit : also note that getfilesizeex is a BOOL returning function. it does not return any information regarding the file size. that is the entire job of the 64 bit int part you are struggling with.

edit2 : also note that the NULL part of getfilesize, if i'm reading this correctly, is actually the leftmost bits that when put together with the returned dword value of getfilesize (the part you have been using on its own) form a 64 bit value in total.... so getfilesizeex is simply combining the 2 together into one data type now that such data types exist in the windows API (whereas previously the largest data type was a single DWORD and so the getfilesize function was designed to split the return value into 2 parts if the file size was exceeded a DWORD).

or, i could be wrong about all of it.
Congratulations ScalaFX for coming out of the closet!
  • #3
  • Edited 6:20pm Sep 9, 2011 3:13pm | Edited 6:20pm
Here is how I'd do it (Note, I have not run the code below. It is cut out of some other code I wrote):

PHP Code
#import "kernel32.dll"
   int  GetFileSizeEx(int FileHandle, int& FileSize[]);
#import

start() {
  int iaFileSize[2];
  double dFileSize;
  int hFileHandle;
  
  hFileHandle=FileOpen("my_big_file.csv",FILE_CSV|FILE_READ,';'); 
  if (hFileHandle>0 && GetFileSizeEx(hFileHandle,iaFileSize)) {
    dFileSize=LIntToDouble(iaFileSize);
    Comment("File size is: "+DoubleToStr(dFileSize,0));
  }
  else {
    Comment("File access error");
  }
  return(0);
}


//--------------------------------------------------------------------------
// Function: LIntToDouble()
// Description: Convert a LARGE_INT to a double
//              LARGE_INT = int ia[2]
//--------------------------------------------------------------------------
double LIntToDouble(int& iaLint[]) {
  if (iaLint[0]<0) 
    return(iaLint[1]*MathPow(2,32)+MathPow(2,31)+(iaLint[0] & 0X7FFFFFFF));
  else 
    return(iaLint[1]*MathPow(2,32)+(iaLint[0] & 0X7FFFFFFF));
}
  • #4
  • Sep 9, 2011 3:30pm Sep 9, 2011 3:30pm
Thanks to both of you for your feedbacks

Xaphod, especially appreciate you posting the example code: will give it a try.

THANKS - MJ
__Thanks__ MJ
  • #5
  • Sep 9, 2011 11:44pm Sep 9, 2011 11:44pm
Quoting Xaphod
Here is how I'd do it (Note, I have not run the code below. It is cut out of some other code I wrote):

[php]
#import "kernel32.dll"
int GetFileSizeEx(int FileHandle, int& FileSize[]);
#import

start() {
int iaFileSize[2];
double dFileSize;
int hFileHandle;

hFileHandle=FileOpen("my_big_file.csv",FILE_CSV|FILE_READ,';');
if (hFileHandle>0 && GetFileSizeEx(hFileHandle,iaFileSize)) {
dFileSize=LIntToDouble(iaFileSize);
Comment("File size is: "+DoubleToStr(dFileSize,0));
}
else {
Comment("File access error");...
that's along the lines of how i would have done it... except, why do you push the result into a double?
Congratulations ScalaFX for coming out of the closet!
  • #6
  • Sep 10, 2011 5:37am Sep 10, 2011 5:37am
Quoting nubcake
that's along the lines of how i would have done it... except, why do you push the result into a double?
I'm quite sure the reason is that the type integer can not hold such large values.

e.g. I tested it on a 10+GB file: exact: 10.7 GB (11452591616 bytes)

if I use the something along the line Xaphod kindly posted: it works exact: 11452591616

if I change it to return int: it works for small files but in the above case I get back: -1432310272
__Thanks__ MJ
  • #7
  • Sep 10, 2011 6:23am Sep 10, 2011 6:23am
yeah that's because int's are signed int's.. so once it reaches a certain number it wraps around to the negative side of the coin.

i didn't realize that doubles in mql could hold such large ranges. i was expecting it to match a signed int plus a fractional component, but it seems doubles in mql hold friggin huge numbers. interesting. didn't realize that, not that it really makes any difference to anything i do.
Congratulations ScalaFX for coming out of the closet!
  • #8
  • Sep 10, 2011 6:32am Sep 10, 2011 6:32am
Quoting nubcake
but it seems doubles in mql hold friggin huge numbers. interesting. didn't realize that, not that it really makes any difference to anything i do.
They implemented it for the account balance
__Thanks__ MJ
  • #9
  • Sep 10, 2011 7:32am Sep 10, 2011 7:32am
Quoting M_j
They implemented it for the account balance
lol, i wasn't thinking about it and actually took that at face value for about 1min... then i was whaaa?

nice one.
Congratulations ScalaFX for coming out of the closet!
  • #10
  • Sep 10, 2011 9:01am Sep 10, 2011 9:01am
Quoting nubcake
except, why do you push the result into a double?
A 64 bit integer does not fit into a 32 bit integer.
  • #11
  • Sep 10, 2011 9:42am Sep 10, 2011 9:42am
Quoting Xaphod
A 64 bit integer does not fit into a 32 bit integer.
yeah, but i didn't realize that double's actually used more bits than an int for the whole number part. i was under the impression that a double was using a signed int plus an unsigned int that was used in magical ways... so to speak. that part of floating point operations was something i haven't really delved into, or at least when i did nothing stuck with me.
Congratulations ScalaFX for coming out of the closet!
  • #12
  • Sep 10, 2011 11:25am Sep 10, 2011 11:25am
Quoting Xaphod
Here is how I'd do it (Note, I have not run the code below. It is cut out of some other code I wrote):

PHP Code
//--------------------------------------------------------------------------
// Function: LIntToDouble()
// Description: Convert a LARGE_INT to a double
//              LARGE_INT = int ia[2]
//--------------------------------------------------------------------------
double LIntToDouble(int& iaLint[]) {
  if (iaLint[0]<0) 
    return(iaLint[1]*MathPow(2,32)+MathPow(2,31)+(iaLint[0] & 0X7FFFFFFF));
  else 
    return(iaLint[1]*MathPow(2,32)+(iaLint[0]..
....
hi Xaphod,
I was wondering if you have a ready function which does the opposite:
DOUBLE to LARGE_INT

e.g. one could use for SetFilePointerEx: liDistanceToMove
http://msdn.microsoft.com/en-us/libr...=VS.85%29.aspx

PHP Code
Syntax
  BOOL WINAPI SetFilePointerEx(  
 __in       HANDLE hFile,   
__in       LARGE_INTEGER liDistanceToMove,  
 __out_opt  PLARGE_INTEGER lpNewFilePointer,   
__in       DWORD dwMoveMethod 
);
__Thanks__ MJ
  • #13
  • Sep 10, 2011 2:53pm Sep 10, 2011 2:53pm
Quoting M_j
hi Xaphod,
I was wondering if you have a ready function which does the opposite:
DOUBLE to LARGE_INT
Not too hard to do:
PHP Code
#property copyright "Copyright © 2010, Xaphod"

//+------------------------------------------------------------------+
//| script program start function                                    |
//+------------------------------------------------------------------+
int start() {
  int iaLargeInt[2];
  double dLargeNum=11452591616.0;

  DoubleToLargeInt(dLargeNum,iaLargeInt);

  Comment("Large Number Before Conversion="+DoubleToStr(dLargNum,0), 
          "\nLarge Number After Conversion="+DoubleToStr(LargeIntToDouble(iaLargeInt),0));

  return(0);
}

// For positive integers only!
void DoubleToLargeInt(double dIn, int& iaLargeInt[]) {
  double dHigh,dLow;
  
  dHigh=MathFloor(dIn/MathPow(2,32));
  dLow=dIn-dHigh*MathPow(2,32);
  iaLargeInt[0]=dLow;
  iaLargeInt[1]=dHigh;
  return(0);
}

double LargeIntToDouble(int& iaLint[]) {
  if (iaLint[0]<0) 
    return(iaLint[1]*MathPow(2,32)+MathPow(2,31)+(iaLint[0] & 0X7FFFFFFF));
  else 
    return(iaLint[1]*MathPow(2,32)+(iaLint[0] & 0X7FFFFFFF));
}
  • #14
  • Sep 10, 2011 3:20pm Sep 10, 2011 3:20pm
Thanks - very kind of you.
__Thanks__ MJ
  • #15
  • Sep 10, 2011 3:34pm Sep 10, 2011 3:34pm
Quoting nubcake
yeah, but i didn't realize that double's actually used more bits than an int for the whole number part. i was under the impression that a double was using a signed int plus an unsigned int that was used in magical ways... so to speak. that part of floating point operations was something i haven't really delved into, or at least when i did nothing stuck with me.
This is a standard double, not a MQL double. A double has a 53 bit mantissa, which means that it can represent exactly any integer which fits in 53 bits (15 decimal digits)

So this will work for files smaller than around 8 PetaBytes , after which you'll start losing precision.
  • #16
  • Sep 10, 2011 3:39pm Sep 10, 2011 3:39pm
Xaphod just a small question.


You write:

// For positive integers only!

PHP Code
// For positive integers only!
void DoubleToLargeInt(double dIn, int& iaLargeInt[]) {
  double dHigh,dLow;
is the comment correct.

I looks like that it does the negative too: at least if I use as input double:

double dLargeNum= -1145259162.0;

I get with the double conversion the same back.

PHP Code

double dLargeNum=-1145259162.0;

Comment("Large Number Before Conversion="+DoubleToStr(dLargeNum,0), 
          "nLarge Number After Conversion="+DoubleToStr(LargeIntToDouble(iaLargeInt),0));
__Thanks__ MJ
  • #17
  • Sep 10, 2011 3:53pm Sep 10, 2011 3:53pm
Quoting M_j
X
You write:

// For positive integers only!
is the comment correct.
I have not thought through and tested how the code works with negative numbers in all possible cases, so I cannot state for certain if it does or not. Maybe it does.
  • #18
  • Sep 10, 2011 5:32pm Sep 10, 2011 5:32pm
Anyone spots were I made my mistake with: SetFilePointerEx
http://msdn.microsoft.com/en-us/libr...=vs.85%29.aspx

Attached: Script: ProblemExample.mq4

Somewhere I messed it up with: SetFilePointerEx
Attached Image
Attached File(s)
File Type: mq4 ProblemExample.mq4   6 KB | 317 downloads
__Thanks__ MJ
  • #19
  • Sep 10, 2011 6:31pm Sep 10, 2011 6:31pm
Quoting M_j
Anyone spots were I made my mistake with: SetFilePointerEx
http://msdn.microsoft.com/en-us/libr...=vs.85%29.aspx

Attached: Script: ProblemExample.mq4

Somewhere I messed it up with: SetFilePointerEx
SetFilePointerEx returns a bool, not a double.

_ArDistanceToMove has not been initialized with data on the first 2 calls of SetFilePointerEx().
Inserted Code
int SetFilePointerEx(int FileHandle,[color=Red] int DistanceToMove[][/color], int & NewFilePointer[], int MoveMethod);
Are you sure you can do that? What exactly is MT4 passing to the function? A value? A pointer?

Why do you not use SetFilePointer() instead. It allows you to pass an integer and a pointer to an integer for the high dword.

Or write a dll to handle your file functions.
  • #20
  • Sep 10, 2011 11:28pm Sep 10, 2011 11:28pm
Quoting Adal
This is a standard double, not a MQL double. A double has a 53 bit mantissa, which means that it can represent exactly any integer which fits in 53 bits (15 decimal digits)

So this will work for files smaller than around 8 PetaBytes , after which you'll start losing precision.
gnarly. i just had to wikipedia double precision to get up to scratch. doubles are not how i assumed they were stored, but then again i always assumed i was wrong anyway.

i suppose that there's the potential for a degree of efficiency in the double type for numbers that don't exceed 32 bits in the mantissa, and after that it becomes a laborious task of shifting bits in and out of dwords for older hardware. maybe i'm overthinking it.
Congratulations ScalaFX for coming out of the closet!
0 traders viewing now