For what it's worth I am also able to use a very similar method to extract a maximum value from a huge table very quickly. When I create my temp view I select only the column that I'm using for determining the maximum value. This speeds up the select Max into varname significantly
Thanks Dennis! > On Dec 7, 2014, at 4:00 PM, Dennis McGrath <[email protected]> wrote: > > It is strange that this kind of update is so slow, but that has been the case > as long as I can remember. > > For tables that have a PK I have resorted to this to get a fast update. > > UPDATE Tablename set… where PKID in (select PKID from tablename where……..) > > I’ll remember your method where this will not work. > > Dennis McGrath > > > From: [email protected] [mailto:[email protected]] On Behalf Of Michael J. > Sinclair > Sent: Saturday, December 06, 2014 11:35 PM > To: RBASE-L Mailing List > Subject: [RBASE-L] - Speeding up my with temporary views > > Hi All, > I have some rbase apps that are running a little bit slower than I would > like. The lines of code that are slow, are updates to some very large tables. > For example, I may want to update a column called flag on every line where > the column client_id = 1234 and there are 1000 lines out of a million lines > that meet that criteria. > Even though I have an index on the column client_id, it still takes a few > seconds. > I created a temporary view, single table, and restricted it to just the rows > with the client_id of the current client. All my changes are now much faster > and appear to be being updated in the source table. > Is the use of a temporary single table view the best choice? > Mike >
