Processing Received Strings

chrisj82

Hi all I am having trouble with a certain system below is an example of the string being received (in netlinx string format)

"02,85,'UNIT-5 ALARM',03"

After the above string there is a whole heap of other junk that I dont want, also the above string wont always be the same, there are other types of alarms ect so the char length will vary.

What I want to do is use 85h as my starting delimeter then store all char's to a variable until 03h is received, is it possible, im sure it is, im just not seeing it.

Any help would be much appreciated, thanks guys

PhreaK

Have a look at the native NetLinx functions 'find_string()' and 'mid_string()'. How to use them is detailed in the NetLinx Keywords Help (found in the help menu of NetLinx Studio). Also, the utility function 'string_get_between()' in the NetLinx Common Libraries may be of use.

DHawthorne

Not to be too picky, but the entire purpose of the 02 is to be the start of your packet. By definition, 02 and 03 are start and end characters. The 85 is probably an identifier for your device, but it's also the ASCII code for the letter 'U' ... and if any of your transmissions have a 'U' in them, your code will mistake it for a new packet and foul your parsing up. 02 and 03 don't have any other use, and aren't going to show up by mistake within the packet.

I love protocols that use a start and end character. First, tokenize your buffer by using something like

sToken=REMOVE_STRING(sBuffer, 3, 1) ;

This will give you a string ending with the 03, and remove it from the original buffer. Chances are more than likely that it's going to start with 02, but if it doesn't, simply do a FIND for the 02, and lop the extra off the token with something like this:

sToken = RIGHT_STRING(sToken, (LENGTH_STRING(sToken) - FIND_STRING(sToken, 2, 1))) ;

That will give you everything after the 02 and including the terminating 03. You can break it down further as necessary.

Spire_Jeff

Another approach similar to Dave's:

//inside the parser.
if(find_string(buffer,"2",1) and  find_string(buffer,"02",1)< find_string(buffer,"3",1)){
   remove_string(buffer,"2",1); 
   msg = remove_string(buffer,"3",1);
}

Just another way to skin the cat

Jeff

vining

Here's a fatter cat, and by fatter I mean alot more code but different.

STRING:
	  {
	  LOCAL_VAR CHAR cRX[1024];
	  
	  cRX = "cRX,DATA.TEXT";
	  SEND_STRING 0,"'Some Dev, Buffer RX [ ',DATA.TEXT,' ] <',ITOA(__LINE__),'>'";
	  while(find_string(cRX,"3",1))//just in case we ever rx multiple complete strings
	       {
	       STACK_VAR INTEGER nFBS;
	       STACK_VAR CHAR cSTR[1024];
	       
	       cSTR = remove_string(cRX,"3",1);
	       nFBS = find_string(cSTR,"2",1))
	       if(nFBS)
		    {
		    get_buffer_string(cSTR,nFBS);//remove up to and including "2"
		    set_length_string(cSTR,length_string(cSTR)-1);//set length string to exclude "3" for sending to parser
		    SEND_STRING 0,"'Some Dev, RX [ ',cSTR,' ] Sending to Parser <',ITOA(__LINE__),'>'";
		    fnParseStr(cSTR);
		    }
	       //else, poof goes the stack, if we find 3 with no preceding 2 it's useless garbage
	       }
	  CANCEL_WAIT 'DUMP_RX';
	  if(length_string(cRX))
	       {
	       WAIT 50 'DUMP_RX' //dump if complete string not received x amount of time, don't keep it! 
		    {
		    SEND_STRING 0,"'Some Dev, Buffer timed out waiting for complete string, Dumping [ ',cRX,' ] <',ITOA(__LINE__),'>'";
		    cRX = '';
		    }
	       }
	  }

I would initially just look for ETX ( 3 ) and then once that's found I would check if a STX also preceded it. Otherwise I'd ignore the RX but dump the content if I don't receive the ETX after a specified time. I almost always use a while just in case more then 1 complete string is ever received and WTF even if it isn't possible there's no real difference between using the if or while in this case.

Plus if you're new to parsing or even if you aren't send_string 0's are your friend. So use them and comment them out when the bugs are worked out.