32 bit byte swap

[mrish]

Hello,

I'm trying to read data from Agilent OSA that output the data in 32 bit number through GPIB interface.

if the bytes are numberd as 0, 1 , 2, and 3 - I need to swap 0 and 3, 1 and 2.

that is: to start with 0/1/2/3 and to end with 3/2/1/0.

can someone help how to do that for each 32 bits?

[crossrulz]

Can you give an example of the data message coming from the instrument?  My initial thought is to just use Unflatten From String with a U32 (or I32) as the data type and the Endianess set to Little Endian.

[mrish]

The data set is simply the trace of the OSA.

1001 numbers such as -45.26 for example.

[mrish]

I attach an image of my proposed solution?

I found split and joing function for high and low byts or word - so I split, swap and rejoin in reverse.

what do you think?

[crossrulz]

Ah, so it is just an array of SGL data coming from the instrument.  Here is what I propose

[mrish]

none of the above works.

I'm very high exponnets numbes (E+8).

Any suggestions?

[Albert.Geven]

Is that floating point constant that you show in your jpg a double or a single?

I can't see from the jpg but it makes a big difference in values.

[crossrulz]

---

@Albert.Geven wrote:

Is that floating point constant that you show in your jpg a double or a single?

I can't see from the jpg but it makes a big difference in values.

---

Let's clarify this a little more.  A SGL (Single Precision Floating Point) uses 32 bits to represent a number.  A DBL (Double Precision Floating Point) uses 64 bits.  So when you convert to your floating point, make sure you are using SGL, just like I did in my snippet earlier.