Assigning an ARRAY of an UDT DATALIST (S7)

BEGIN
NEW_FLANK:= NEW AND (NOT MEM_NEW);
DEL_FLANK:= DEL AND (NOT MEM_DEL);
ACT_FLANK:= ACT AND (NOT MEM_ACT); 
IF NEW_FLANK THEN							 // HANDLING NEW BUILDED BATCH
	FOR i:= 1 TO 99 DO
		LIST [101-i]:= LIST[100-i];
	END_FOR;
	FOR j:= 1 TO 4 DO
		LIST [6-j]:= BUILD[5-j];
	END_FOR;	
BATCHES_IN_LIST:= BATCHES_IN_LIST + 5  
END_IF; 
   
IF ACT_FLANK THEN							 // HANDLING ACTIVATION OF A BATCH
  FOR k:= 1 TO 5 DO
		ACTIVE [6-k]:= LIST[BATCHES_IN_LIST+k];
  END_FOR;
  FOR k:= 1 TO 5 DO
		LIST[BATCHES_IN_LIST+k]:= 0;		
  END_FOR;
  BATCHES_IN_LIST:= BATCHES_IN_LIST - 5 
END_IF;	
 
MEM_NEW:= NEW;
MEM_DEL:= DEL;
MEM_ACT:= ACT;
END_FUNCTION_BLOCK

JesperMP said:

Oh, by the way.

In case that the problem is that you cannot get STEP7 to accept that you remove the square brackets, then the reason is that the input_output of the FC and the DB list is formatted differently.
If succesful, "BATCH_LIST".LIST will be accepted, and you can see that it is really a pointer, not an ANY pointer. So STEP7 seems to check both data types and length.

Click to expand...

JesperMP said:

Hi combo.

I am not ging to try and understand your code.
I dont have the time, and even if I had, I think you have some bad coding habits that makes your code almost unreadable.
First, please dont write variables with capitals only !
And somehow I think there is some structure behind that you have different code for different areas of the arrays, but it is impossible to know for an outsider. I would create the structure that matches these differencs, and not hide it behind some constants you need to know.

For example, to push the list one step forward, I would do:

FOR index:= NoOfListEntries TO 2 BY -1 DO

"BATCH".List[index].Number := "BATCH".LIST[index-1].Number ; 

"BATCH".List[index].Length := "BATCH".LIST[index-1].Length ; 

"BATCH".List[index].Width := "BATCH".LIST[index-1].Width ; 

END_FOR;

This is simpler and much more readable.
It is also more maintainable. If you change the structure, by adding a new parameter for example, the code only has to be modified by adding the new parameter, and you dont have to know how to change some constants to some other constants.

Click to expand...

JesperMP said:

With the FOR loop you can specify the start, end and step.
So assuming you want to do something with the last entry and move backwards to the start, you have to step one entry backwards for each loop. That accounts for the "-1". The rest I think you can figure out easily.
edit: In your own code there is some nasty programming that does the same I think:

    FOR i:= 1 TO 95 DO

        LIST [101 - i]:= LIST[96 - i];

    END_FOR;

    FOR j:= 1 TO 5 DO

        LIST [6 - j]:= BUILD[6 - j];

    END_FOR;

.Number, .length and .width are just examples on the structure you can manipulate.
My code assumes that you have made a DB with a UDT like we described before.

Example:

UDT "List_entry" is
.Number DINT
.Length REAL
.Width REAL

DB "BATCH" is
.EDIT "List_entry"
.ACTIVE "List_entry"
.REPORT "List_entry"
.LIST ARRAY[1..99] of "List_entry".

My point is that the code gets more readable if you define structures symbolically.

edit: My code is not an exact solution to your problem. It is merely an example of handling datas in SCL. But I think you can get the idea to write easily understandable code that does what you want.

Click to expand...

JesperMP said:

SFC20 BLKMOV is not really useful to manipulate lists, because "The source and destination areas must not overlap" as it says.

Click to expand...

S7Guy said:

This applies to arrays of udts as well, so life can be a lot simpler if things like that are taken into consideration at the beginning. And, paramterizing the required ANY pointers dynamically is pretty easy too as long as the code is well organized.

Click to expand...