Re: Subfile question

Hello Brian,

Am 17.07.2024 um 11:57 schrieb Brian Parkins <goodprophet.bp@xxxxxxxxx>:

Firstly, remember the information loaded in the subfile is a duplicate of the source data. As time passes, this duplicate will become increasingly "stale" and no longer represent the accuracy of the underlying data. Depending on the application, this may not be an issue. However, if currency/accuracy IS important, then a single-page subfile, (SFLPAG=SFLSIZ) may be more appropriate.

This is one of the reasons I prefer load-page over load-all SFLs.

Also, I found it way easier to implement position-to compared to load-all. Also, the user is able to scroll backwards from the record he was presented with pos-to due to "relative reading". Another nice feature about my particular implementation is that meanwhile deleted or added records are automatically recognized during scrolling and pos-to.

https://github.com/PoC-dev/as400-sfltemplates

Secondly, if sensible filtering is in place the User should not be paging up/down more than a couple of screens-worth. So why load an entire subfile with 100's of rows? Again, depends on the application and how much data is involved.

Depends: Yes. In general, I try to avoid the requirement to scroll more than four to five pages. Of course, this is only possible when the eventual amount of records is already known.

Since the maximum size of a subfile is only 9,999 rows, I don't think performance was ever an issue. IMHO, the three different "load techniques" address different usability criteria - i.e. one size does not fit all situations.

I disagree. Loading 500 records into a load-all SFL is much slower on my 150 than loading just 13 into a load-paged. Even more on my P03. *After* the loading, scrolling is faster, though. The difference is most likely negligible on machines from the current decade.

As you say, it depends on the application. :-)

:wq! PoC