PARTITION (
$D15.APDATA.STUFF02
CATALOG \NODE.$D01.APCAT
EXTENT(40000, 80000)
FIRST KEY (<some very large value>)
etc ...
To keep rows from "bleeding" into the empty first partition, they should
have the theoretical maximum value of the first column in that FIRST KEY ()
clause, correct? Not the run of the mill lowest value for a primary key
column in "regular old" ASCending order. That's how I'm reading the SQL/MP
Reference Manual, but it may be tricking my tiny brain into delusion.
Thanks much,
Brian Ray
PARTITION 0 (ANCHOR): FIRST KEY (0)
PARTITION 1: FIRST KEY (1)
PARTITION 2: FIRST KEY (2)
PARTITION 3: FIRST KEY (3)
PARTITION 4: FIRST KEY (4)
PARTITION 5: FIRST KEY (5)
PARTITION 6: FIRST KEY (6)
PARTITION 7: FIRST KEY (7)
PARTITION 8: FIRST KEY (8)
PARTITION 9: FIRST KEY (9)
Since he says the index is descending, is your answer correct?
Brian,
SQL/MP Reference Manual - Chapter P, heading "Partition Clause" (page
P-17 in the G06.28 online version).
Under the discussion of FIRST KEY parameter, it says FIRST KEY
"specifies the lowest value for the partition if the column for the
value has an ascending collating sequence; it specifies the highest
value for the partition if the column has a descending collating
sequence."
On Jul 20, 3:31 am, Keith Dick <kdic...@comcast.net> wrote: