in switch stage I have selected selectore mode as hash the documentation says following
"Hash means that rows are hashed on the selector column modulo the number of output links and assigned to an output link accordingly"
I am not getting what it does
my job has 3 output link and selectore column has values 2,3,10, 1
then in Out put I should get
2mod3 - assign to output dataste with label 2
3mod3 - assign to output dataste with label 0
10mod3 - assign to output dataste with label 1
1mod3 -assign to output dataste with label 1
is my understanding is correct
Switch
Moderators: chulett, rschirm, roy
LegendKiller,
you are correct, the modulo goes from link number 0 to {n-1}.
you are correct, the modulo goes from link number 0 to {n-1}.
<a href=http://www.worldcommunitygrid.org/team/ ... TZ9H4CGVP1 target="WCGWin">
</a>
</a>
-
- Participant
- Posts: 54607
- Joined: Wed Oct 23, 2002 10:52 pm
- Location: Sydney, Australia
- Contact:
-
- Participant
- Posts: 60
- Joined: Sun Nov 21, 2004 2:24 am
ray.wurlod wrote:Where DO the rows go, given that you seem to be implying that they don't go where you expect them to?
yes in my case
when selector value are 2,1 then they are going to dataset label 2
when selector value are 3,10 then they are going to dataset label 1
so it is not behaving as I expected
the input selector value are 2,3,10,1 and outpit link are 3 so i expected that
row containing value 2 will go to dataset 2
row containing value 3 will go to dataset 0
row containing value 1,10 will go to dataset 1
do i need to put partition as hash
-
- Participant
- Posts: 54607
- Joined: Wed Oct 23, 2002 10:52 pm
- Location: Sydney, Australia
- Contact:
Your expectations are based on a modulus algorithm, so why not use that? You can not assert that a hashing algorithm will produce the same results as a modulus algorithm - hashing algorithms incorporate extra processing to try to guarantee repeatably-random but even allocation.
IBM Software Services Group
Any contribution to this forum is my own opinion and does not necessarily reflect any position that IBM may hold.
Any contribution to this forum is my own opinion and does not necessarily reflect any position that IBM may hold.
-
- Premium Member
- Posts: 68
- Joined: Sat Feb 05, 2005 4:59 am
- Location: India
Hilegendkiller wrote:ray.wurlod wrote:Where DO the rows go, given that you seem to be implying that they don't go where you expect them to?
yes in my case
when selector value are 2,1 then they are going to dataset label 2
when selector value are 3,10 then they are going to dataset label 1
so it is not behaving as I expected
the input selector value are 2,3,10,1 and outpit link are 3 so i expected that
row containing value 2 will go to dataset 2
row containing value 3 will go to dataset 0
row containing value 1,10 will go to dataset 1
do i need to put partition as hash
Have u been able to sort the anomaly
Have been trying to simulate the same scenario and am hitting the same prblm even after using modulo as the partitioning algo.
Thnx