Stas: This is timely since just this week we had a big discussion about
changing pointers to fetch statements!
There are multiple ways to design a declare cursor. Here's how I would do
it. The "while 1 = 1" is kind of a dummy statement that will always
evaluate as true. The "IF Sqlcode" that follows the fetch is what will break
out
of the while loop as soon as there is no more data to process. I notice
you didn't have an "open" in your example so you need to have that in there.
DECLARE cursor1 CURSOR FOR SELECT itemnumber, +
iccostlbs, icvalue, inweight, datereceived, inhouse +
FROM icmain WHERE inhouse CONTAINS 'Y' +
ORDER BY itemnumber ASC
OPEN cursor1
WHILE 1 = 1 THEN
FETCH cursor1 INTO itemnumber, iccostlbs, +
icvalue, inweight, datereceived, inhouse
IF SQLCODE = 100 THEN
BREAK
ENDIF
-- do your processing here
ENDWHILE
Karen
> Stas' Gawel [[email protected]]
> Tube Methods, Inc.
> 610-279-7700
>
> I am converting an rbase command file from an older version of rbase to
> rb8 but I do not know how to completely convert the pointer command.
> This is the old syntax involving the pointer command
>
> SET POINTER #1 STAT1 FOR ICMAIN WHERE +
> INHOUSE = "Y" SORTED BY ITEM#
> WHILE STAT1 = 0 THEN
> SET VAR ITEM# = ITEM# IN #1
> SET VAR COSTLBS = COSTLBS IN #1
> SET VAR VALUE = VALUE IN #1
> SET VAR INWEIGHT = INWEIGHT IN #1
> SET VAR DATERCVD = DATERCVD IN #1
> SET VAR INHOUSE = INHOUSE IN #1
> NEXT #1 STAT1
> ENDWHILE
>
> This is as far as I got.
>
> DECLARE cursor1 CURSOR FOR SELECT itemnumber, + iccostlbs,
> icvalue, inweight, datereceived, inhouse +
> FROM icmain WHERE inhouse CONTAINS 'Y' +
> ORDER BY itemnumber ASC
> WHILE STAT1 = 0 THEN
> FETCH FROM cursor1 INTO itemnumber, iccostlbs, +
> icvalue, inweight, datereceived, inhouse
> ENDWHILE
>
> I am having trouble with what STAT1 is and what it should be converted
> to.
> Also, I receive the error message -ERROR- The cursor [cursor1] is
> already defined. (708) but I have checked and my syntax appears to be
> right for the declare and fetch commands.
>
> Note that some of the coulumn names changed from the older version of
> the table and the newer one though the table name is still the same and
> that this is not the whole program.
>
