| |
|
RSS Feed - WebSphere MQ Support
|
RSS Feed - Message Broker Support
|
  |
|
| MOVE NEXTSIBLING vs CARDINALITY and [i] call in a loop |
« View previous topic :: View next topic » |
| Author |
Message
|
| Bartez75 |
 Posted: Mon Jun 08, 2009 4:04 am Post subject: MOVE NEXTSIBLING vs CARDINALITY and [i] call in a loop |
 |
|
Voyager
Joined: 26 Oct 2006
Posts: 80
Location: Poland, Wroclaw
|
Hi
Could anybody tell me what is better to use when looping through message. In my case input message is MRM/XML. Output is MRM/TDS.
There is let say 1000 of Order element in Data.
I can do a loop like that:
| Code: |
DECLARE i INTEGER 1;
DECLARE refOrder REFERENCE TO InputRoot.MRM.ns:Data.ns:Order[1];
WHILE LASTMOVE(refOrder) DO
SET OutputRoot.MRM.Order[i].NMCIL=refOrderCIL.id;
SET OutputRoot.MRM.Order[i].NUBSL=refOrderCIL.ns:Row.slRow;
MOVE refOrder NEXTSIBLING NAMESPACE * NAME 'Order';
SET i=i+1;
END WHILE;
|
or I can do it like that:
| Code: |
DECLARE i INTEGER 1;
DECLARE orderCardinality INTEGER CARDINALITY(InputRoot.MRM.ns:Data.ns:Order[]);
WHILE i <= orderCardinality DO
SET OutputRoot.MRM.Order[i].NMCIL=InputRoot.MRM.ns:Data.ns:Order[i].id;
SET OutputRoot.MRM.Order[i].NUBSL=InputRoot.MRM.ns:Data.ns:Order[i].ns:Row.sRow;
SET i = i + 1;
END WHILE; |
There are more fields to map. I just put these 2 to make this example simple.
From my tests the first one (with MOVE NEXTSIBLING) is faster but maybe my test was wrong. Is MOVE NEXTSIBLING a better way of doing it? |
|
| Back to top |
|
 |
| Vitor |
| Posted: Mon Jun 08, 2009 4:12 am Post subject: Re: MOVE NEXTSIBLING vs CARDINALITY and [i] call in a loop |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| Bartez75 wrote: |
| Is MOVE NEXTSIBLING a better way of doing it? |
I would expect this to be the faster method, but that's based on my (possibly flawed) understanding of how the tree would be navigated.
I'm sure someone with a more authoritative view will be along in a minute.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| mqjeff |
| Posted: Mon Jun 08, 2009 4:22 am Post subject: |
|
|
Grand Master
Joined: 25 Jun 2008
Posts: 17447
|
It should be long established that using references to navigate trees is better performing.
In order to find the [n+1]'th entry when using indices, Broker must navigate to [1], [2], [3],....[n] and then to [n+1].
When using a reference, it just accesses the relevant pointer stored in the node datastructure. |
|
| Back to top |
|
|
| Vitor |
| Posted: Mon Jun 08, 2009 4:44 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| mqjeff wrote: |
In order to find the [n+1]'th entry when using indices, Broker must navigate to [1], [2], [3],....[n] and then to [n+1].
When using a reference, it just accesses the relevant pointer stored in the node datastructure. |
I'm relieved my understanding is not as flawed as I feared. 
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| Bartez75 |
| Posted: Tue Jun 09, 2009 3:16 am Post subject: |
|
|
Voyager
Joined: 26 Oct 2006
Posts: 80
Location: Poland, Wroclaw
|
Thanks for you answer.
I will be closer with references after your explanation  |
|
| Back to top |
|
|
|
|
|
| |
|
Page 1 of 1 |
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
|
|
