Re: Sorting and filtering subfiles

[Dave Murvin <davem@xxxxxxxx>]

Don't know if this is better or faster, but this is how I have done it for 
a few applications.

1. Load all the info to be displayed in the subfile into an array (multiple 
fields in an array described in a data structure), expanding the array as 
required using alloc and realloc.  When I am finished filling the array, I 
do one more re-alloc to allocate memory equal to the number of records in 
the array.
2. Sort the array in memory in the default subfile sequence using qsort 
(this might not be necessary if you initially load the array in the default 
sequence).
3. Load the subfile from the array.  Either the whole subfile or page at a 
time.
4. When user selects a different sequence, resort the array in memory using 
qsort and reload the subfile.

I know this may not work for a huge file, but with a limit of 9,999 records 
in a subfile (unless you use page at a time) and for my applications, this 
seems to work okay.  I originally got the qsort sample from Jon's "Who Knew 
You Could Do that with RPGIV?" book.

I have thought about using the double click method of letting them sort the 
subfile, but I thought that not everyone would know to double click the 
columns.  What I have done so far is to add a function key for "Sort 
Display" then give them a pop up window that allows them to select the 
sequence.    The pop up window also gave me a change to try out the 
SNGCHCFLD (Single-Choice Selection Field) keyword on the display.

At 03:19 PM 3/29/2005, you wrote:
> I write out the subfile to a work file in qtemp,
> which has the same fields with one extra "KEY" field.  I load the sorted
> columns data into both the matching field and the "KEY" field, then reload
> the subfile from the indexed work file in QTEMP.    I am sure there are fast
> and better ways,...

Dave Murvin
DRM Enterprises, Inc.