Forum Discussion
Most Efficient Method to Filter on Progress Database
I like option 3 best as long as your list is < 1,000 items. More than that and it starts to get slow.
Because you are not folding, you have to process 100% of the data no matter what.
If Table 1 isn't huge after the date filter, consider wrapping it in Table.Buffer before you do the merge. Test that out - but save the file first. If it is huge, it could eat up RAM and cause the app to lock up. I speak from experience when I used Table.Buffer on a 3M record 30+ column table.
Thanks Edhans.
Yea most of the time it isn't too bad (I consider report refresh in sub 2 minutes acceptable for users) but as you say utilising bigger tables with 2M+ lines things start to slow down considerably.
I've played with Tabe.Buffer a bit to see if that can speed things up but it seems really hit and miss either getting a slight speed increase, no change or absolutley kills the report depending on where I use it.
Main speed considerations/rules I try to stick to:
- Filter as early as possible
- Remove unwanted columns before merging
- As few steps as possible and no duplication of steps (so all my change types, renames etc. together)
- Reference tables rather than going into the databse multiple times for the same data
Are there any other 'speed' guidelines people try and use when query folding isn't an option?