Can VFP read DOUBLEs?

[chpicker]

Is there a way to read binary file formats in VFP? For instance, I have a file that contains 8-Byte DOUBLEs. I need to read these in and store the numbers they represent in a table. For instance:
[tt]
82 03 5A BA 82 E6 57 C0 = -95.6017
9C F8 6A 47 71 B4 47 40 = 47.4097
[/tt]
Is there any way to do this in Fox?

Ian

 

[wgcs]

try FOPEN() and FREAD()... It won't automatically handle type conversions, but it will give you the info as Characters that you can convert and manipulate with ASC(), BITRSHIFT() and BITLSHIFT()

 

[chpicker]

Ugh...no, that won't work. I do that already with Integers, but when it comes to Doubles...

The formula for a DOUBLE is:

1 bit: sign (0 is positive, 1 is negative)
11 bits: Exponent (power of 2)
52 bits: Mantissa (fraction expressed as a power of 2 with an implied 1 added)

You multiply the 3 parts together to get the final number.

The part that I just can't figure out is how to calculate that Mantissa. In any case, it's really not worth the effort. I think it would be easier for me to learn how to create a DLL in C++ that will do it for me (C++ can do automatic type conversions).

Assuming no one has any other suggestions, I'll see if I can figure out how to do that instead.

Ian

 

[wgcs]

My first answer was just a quickee... I've been looking up just what you've found... the conversion is the pain.

This page (at the bottom) has all the formulas for a Single, which shouldn't be hard to extrapolate to a double:

http://www.cs.bc.edu/~straubin/160notes/l2/lect2.html

The one statement on that page I'm still trying to understand is "The exponent field holds the binary representation of 127, and thus the exponent is 0."

 

[rgbean]

Ah, I understand that part. While the first bit is the sign of the "whole" number, the exponent can also be positive or negative. So, when the exponent is 0 -> 126 the value is negative (neg + 127), for 127 it's 0, and 128 -> 255 it's positive (minus 127). e.g. -3 exponent is 124, and +3 is 130. Of course those values are again for a Single (w/ 8 bit exponent), they'd need to be a adjusted appropriately for a Double with it's 11 bit exponent!

Rick

 

[wgcs]

Ah! that makes sense!

 

[wgcs]

Here's a solution, finally! Thanks to PeterEasson:

http://fox.wikis.com/wc.dll?Wiki~VFPFloatingPointDataType

 

[738262]

Hi,
See this Faq I created.
faq 184-1932

 

[ChrisRChamberlain]

738262

The syntax faq184-1932 provides a link HTH

Chris

 

[chpicker]

I actually ended up creating an FLL file in Visual C++ that does the conversion for me. The code is very small and very fast. It receives an 8 byte string, writes it directly to memory, and reads the memory as a DOUBLE value. Then it returns the DOUBLE that it sees in that memory location.

I can post the code if anyone is interested. I would just do it, but the code is at work, and I don't go back in until Monday. LOL

Of course, I could also just post the FLL file.

Ian