Is the internal methods behind REMOVE the same as EXTRACT (or <>). Again, about a year ago, I had an issue with an attributed xref rec of the invoices for a customer that had grown to almost 150,000 invoices.
The original programmer used <> with a counter and it took very long to 'age' that cust on an inquiry screen. (don't solve my A/R problems). This was on a D3 box. So I utilized the REMOVE command and it went from around 3 minutes to around 15 seconds. BTW. To use REMOVE on D3, you need the OPTION $EXT command earlier. I'm gonna try your LEN vs #"" to see if that makes sense. As an experienced programmer though, I would have to see if it really matters on regular small values instead of these monster values. I've used #"" a zillion times. Thanks Mark johnson ----- Original Message ----- From: "Stevenson, Charles" <[EMAIL PROTECTED]> To: <u2-users@listserver.u2ug.org> Sent: Friday, March 09, 2007 1:48 AM Subject: RE: [U2] old dog, new tricks > To add to Adrian's explanation, when a string is in memory, the first > few bytes at that string's address is not actual data, but metadata > about the string, including its length, so UV immediately knows where > the string ends (& where the next string begins). > Therefore <-1> can make an immediate jump, rather than searching through > the string counting attributes. > > This also implies that something like > IF LEN( X ) THEN > is faster than something like > IF (X # '') THEN > . > > You can also see why <n,-1> & <n,m,-1> aren't particularly fast, since > they aren't at the end of the srting. > > UV keeps similar metadata up front about the last attribute accessed, & > the remove pointer. That is what makes both REMOVE/REVREMOVE and > EXTRACTs of sequential <n>s so fast. > > cds > ------- > u2-users mailing list > u2-users@listserver.u2ug.org > To unsubscribe please visit http://listserver.u2ug.org/ ------- u2-users mailing list u2-users@listserver.u2ug.org To unsubscribe please visit http://listserver.u2ug.org/
