I did implement what I described below, which I believed to work well
unless it is the first row that needs the biggest buffer. However, I
ran into a problem. I keep getting incorrect results when I use the
coalesce() function. Usually coalesce() will simply return the first
argument, and if
On 02/24/2012 02:02 PM, Steinar Midtskogen wrote:
Hello
Is it possible to find out in xFilter or xBestIndex which columns were
selected? That is, if I do SELECT a, b, c FROM t where t is a
virtual table, I would like to know in xFilter or xBestIndex that the
result will only consist of the
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 23/02/12 23:02, Steinar Midtskogen wrote:
I know that xColumn will only get called for these columns.
As Dan said there isn't a direct way of knowing. There is a reasonably
convenient workaround of having multiple virtual tables with different
Thanks to Dan and Roger for the information and suggestions. I have
about 100 columns (and millions of rows), but usually queries will
only ask for a few columns.
I think I know a way to work around this for my case. Basically, I
don't know the exact size required for my lookahead buffer
Hello
Is it possible to find out in xFilter or xBestIndex which columns were
selected? That is, if I do SELECT a, b, c FROM t where t is a
virtual table, I would like to know in xFilter or xBestIndex that the
result will only consist of the rows a, b and c.
I know that xColumn will only get