Thanks Larry,

The database is an index of board of trustee's meetings.
It is essentially an outline.

the columns are:

[heading]               [subheading]    [subsubheading] [date]  [other]

The headings, subheading etc are reused often. Also in the entry form the client wants lookups for the 3 heading and the subheading based on the heading and the subsubheading based on the subheading.

Would I need to create 3 unique indexes for the 3 seperate combo boxes?

CREATE UNIQUE INDEX ON board_trustees(heading)
CREATE UNIQUE INDEX ON board_trustees(heading,subheading)
CREATE UNIQUE INDEX ON board_trustees(heading,subheading,subsubheading)

???

James

On Jul 16, 2004, at 4:25 PM, Lawrence Lustig wrote:

Thanks, however I tried the temp table thing already
and that takes too
long also. Its creating a temp table then inserting
non-duplicate rows
from a table that has 14,000 records in it and that
was too painfull
too

If you are trying to dedup a 14K record table everytime you load the combo box, you may have a legitimately slow operation on your hands. Certainly, the combo box is unlikely to load _faster_ than the time it would take to select the records into a temp table (the problem is the combo boxes are often much _slower_ than a regular SELECT).

But here's something you can try.  Create a compound
index on your table including all the columns that
appear in the combo box and that have to be unique.
If you do that, R:Base _should_ perform and index-only
retrieval for the combo box, and that will have
performed the deduplication for you:

CREATE UNIQUE INDEX ON YourTable(YourCol1, YourCol2,
YourCol3)

--
Larry

James Hageman, CITA I
University of Delaware Archives
002 Pearson Hall
Newark ,DE 19702
302-831-3127
[EMAIL PROTECTED]



Reply via email to