| Author |
Message
|
| cloucas |
 Posted: Thu May 04, 2006 5:15 am Post subject: Referencing multiple instances of an element in ESQL |
 |
|
Apprentice
Joined: 10 Mar 2006
Posts: 37
|
I have a message set that conatins a message whose type consists of a header and a body. The body is a complex type and is composed of three different complex types, such that:
[Header][[Type1][Type2][Type3]]
Type2 above is a field delimited type that can have one or more occurences of a pair of elements, such that:
[AccountNo][Balance], each element is ofcourse preceded by its delimiter.
So for instance, if an account holder has three accounts, the message would look like:
[Header][[Type1][(Account1)(Balance1)(Account2)(Balance2)(Account3)(Balance3)][Type3]]
The number of accounts is different from accountholder to accountholder and is thus not fixed.
My question is how to reference each of these instances of accounts in the message from my ESQL (i.e. is there an 'array-like' reference I can use?).
Thanks in advance
Chris |
|
| Back to top |
|
 |
| elvis_gn |
| Posted: Thu May 04, 2006 5:22 am Post subject: |
|
|
Padawan
Joined: 08 Oct 2004
Posts: 1905
Location: Dubai
|
Hi cloucas,
Well described problem 
You should search for REFERENCE TO, MOVE, NEXTSIBLING...search in the forum or in the ESQL pdf.
Regards. |
|
| Back to top |
|
|
| jefflowrey |
| Posted: Thu May 04, 2006 5:29 am Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002
Posts: 19981
|
In fact, it's not an "array-like" syntax. It's so much exactly like an array syntax that people get confused and think it IS an array.
Set OutputRoot.MRM.Type1.Account[3] = "Account3";
_________________
I am *not* the model of the modern major general. |
|
| Back to top |
|
|
| sirsi |
| Posted: Thu May 04, 2006 7:54 am Post subject: |
|
|
Disciple
Joined: 11 Mar 2005
Posts: 177
|
shouldnt this be
Set OutputRoot.MRM.Type2.Account[3] = "Account3";
instead of
Set OutputRoot.MRM.Type1.Account[3] = "Account3";
considering Account details occur inside Type2 |
|
| Back to top |
|
|
| cloucas |
| Posted: Thu May 04, 2006 8:51 pm Post subject: |
|
|
Apprentice
Joined: 10 Mar 2006
Posts: 37
|
OK Guys
Thanks very much for your help....any ideas on how to reference the number of elements in the array? My idea is to loop through all elements...is there a "FOR EACH...." loop in ESQL or must I loop through using FOR loop?
Thanks
Chris |
|
| Back to top |
|
|
| elvis_gn |
| Posted: Thu May 04, 2006 9:12 pm Post subject: |
|
|
Padawan
Joined: 08 Oct 2004
Posts: 1905
Location: Dubai
|
Hi cloucas,
Use what ever you can get it working with...dont your FOR else u'll have to find an integer i and store all available child elements etc....
U can use a WHILE LASTMOVE(reference) with a MOVE reference NEXTSIBLING....
Regards. |
|
| Back to top |
|
|
| fjb_saper |
| Posted: Fri May 05, 2006 2:32 am Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
The right way to do this would be to change your message set definition and create a header level:
MRM-type1
-type2 accounts(group level)
-account(=subtype 2) ,0, unbounded
-type3
And address them by OutputRoot.MRM.type2.*[]
This way you can do a cardinality on them and really treat them as an array.
Enjoy 
_________________
MQ & Broker admin |
|
| Back to top |
|
|
| kimbert |
| Posted: Fri May 05, 2006 6:44 am Post subject: |
|
|
Jedi Council
Joined: 29 Jul 2003
Posts: 5542
Location: Southampton
|
fjb_saper: Not sure what you mean.
- cloucas's logical model sounds fine to me - why should he change it?
- CARDINALITY works on any node in the message tree - it just returns the number of child nodes. This is independent of the message set ( if any ) which the parser used to produce the tree.
- your syntax 'OutputRoot.MRM.type2.*[]' looks strange. Why the anonymous field reference? Surely cloucas wants paths like 'Set OutputRoot.MRM.Type2.Account[3]' and 'Set OutputRoot.MRM.Type2.Balance[3]'
I've probably misunderstood what you were saying, but I'm posting this in case anyone else gets confused. |
|
| Back to top |
|
|
| sirsi |
| Posted: Fri May 05, 2006 7:17 am Post subject: |
|
|
Disciple
Joined: 11 Mar 2005
Posts: 177
|
| i have another question... which is the most efficient way of looping through multiple occurrences... i remember reading somewhere that using "WHILE LASTMOVE(reference)" method should be the efficient one... |
|
| Back to top |
|
|
| jefflowrey |
| Posted: Fri May 05, 2006 8:35 am Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002
Posts: 19981
|
| sirsi wrote: |
| I remember reading somewhere that using "WHILE LASTMOVE(reference)" method should be the efficient one... |
It certainly should be. If it isn't, you can probably open a PMR.
_________________
I am *not* the model of the modern major general. |
|
| Back to top |
|
|
| fjb_saper |
| Posted: Fri May 05, 2006 9:01 pm Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
| kimbert wrote: |
fjb_saper: Not sure what you mean.
- cloucas's logical model sounds fine to me - why should he change it?
- CARDINALITY works on any node in the message tree - it just returns the number of child nodes. This is independent of the message set ( if any ) which the parser used to produce the tree.
- your syntax 'OutputRoot.MRM.type2.*[]' looks strange. Why the anonymous field reference? Surely cloucas wants paths like 'Set OutputRoot.MRM.Type2.Account[3]' and 'Set OutputRoot.MRM.Type2.Balance[3]'
I've probably misunderstood what you were saying, but I'm posting this in case anyone else gets confused. |
The problem from what I understand is the way he defined his message set.
By not defining a header level for his occurs he actually gets a flat recurrence:
- type1
- type2
- type2
- type2 n
- type 3
This makes manipulations very difficult as it is not easy to determine the number of occurrences of type2
By defining his structure a little bit differently (the data on the wire may still look the same) he gets some advantage
-type1
-type2 - child(0, unbounded)
-type3
and can access it directly in a true array style :
OutputRoot.MRM.type2.child[n]
Declare k integer cardinality (OutputRoot.MRM.type2.*[]);
Otherwise there is a hard time getting the number of occurrences as
declare k integer cardinality(OutputRoot.MRM.*[]) will also return type1 and type3 occurrences...
Of course you could try a select to determine the cardinality but if you have to manipulate the data the other way makes it just plain more simple. (XML rule -- always have a header level for the items of a collection...)
Hope this clarifies a bit
Enjoy
_________________
MQ & Broker admin |
|
| Back to top |
|
|
|
|
