Re: Regarding string as index for tables

Dave,

Thanks.

Now that this should help me out on the double index issue..

1) If the table is like this and the none mentioned for second index(so assuming non- implied...)

Index1   Index2  c1  RS
---------------------------------- 
AB          BEA      1
AB          DC      2
BC          QF      30
DE          FC      15
DE          AD      10

Then walk should be like  

c1.ABDC 
c1.ABBEA 
c1.BCQF 
c1.DEAD
c1.DEFC

 And if second index is IMPLIED then the walk is :

c1.ABBEA 

c1.ABDC 

c1.BCQF 

c1.DEAD

c1.DEFC


Am I right?

2)You mentioned 'iterator' helper. Is this generic library helper or any specific one? 
Can you provide some examples pl.?

3)If the example from my previous post is altered to 


> index  c1  rowstatus
> ----------------------
> ab     1
> ag     1  <<<  this is new addition
> hi     2
> cd     3
> ef     4
> bc     5

In that case, the table **as reported via SNMP** should be walked in the order

    c1."ab" = 1
    c1."ag" = 1 <<<< Here goes the new one ..as 'a' in ab and 'a' in ag are same then next comparision is between b and g so g comes second as per lexicographis order am I right?
    c1."bc" = 5
    c1."cd" = 3
    c1."ef" = 4
    c1."hi" = 2
    rowstatus."ab" = ...
    rowstatus."bc" = ...
    rowstatus."cd" = ...
    rowstatus."ef" = ...
    rowstatus."hi" = ...

And again if this is modified to 

> index  c1  rowstatus

> ----------------------
> ab     1

> abc       1 <<< new one and single letter 

> ag     1 

> hi      2

> cd     3

> ef      4

> bc     5

and assuming that the index is non-IMPLIED which one comes first in the walk 
c1.ab or c1.ag ? as both got 2 letters ? May I assume this is going to be implementation specific as some or other logic needs to be applied at the end to retrieve the data for non-implied index....and 'abg' comes last in the walk as this is three letter


Regards
--- On Thu, 15/4/10, Dave Shield <[email protected]> wrote:

From: Dave Shield <[email protected]>
Subject: Re: Regarding string as index for tables
To: "phani kumar" <[email protected]>
Cc: [email protected]
Date: Thursday, 15 April, 2010, 9:50

On 15 April 2010 09:34, phani kumar <[email protected]> wrote:
> Thanks for quick response and as you said lets assume the index looks like
>
>  { index1,   IMPLIED index2 }
>
> and lets simplify the index from two to one

Which?
Are you asking about a two-string indexed table, or a single-string
indexed table?


> and lets say this looks like
>
> index  c1  rowstatus
> ----------------------
> ab     1
> hi     2
> cd     3
> ef     4
> bc     5

In that case, the table **as reported via SNMP** should be walked in the order

    c1."ab" = 1
    c1."bc" = 5
    c1."cd" = 3
    c1."ef" = 4
    c1."hi" = 2
    rowstatus."ab" = ...
    rowstatus."bc" = ...
    rowstatus."cd" = ...
    rowstatus."ef" = ...
    rowstatus."hi" = ...



> In the above scenario, the index is not according to lexicographical
> (am i correct?) order and this is because its is not sorted(lets assume).

I presume this is how the data is represented internally?
(Either within the agent, or in the underlying subsystem).

In which case, one of the tasks of the MIB implementation module
is to present the data in lexicographical order, as above.
That's precisely the purpose of the 'iterator' helper.


>      So, how does the getnext works here?
> How is the next lexicographically index found? Is this like this
>
> ab then bc then  cd then ef and hi

Correct.

Note that this works "as expected" because all the strings are
the same length.   If some of the index strings were longer
than others, then the order depends on whether the table is
IMPLIED or not.
   An IMPLIED index would continue to work in dictionary order.
A non-IMPLIED index would sort by length first.  So short index
strings would come before longer ones.

OK?

Dave

> Could you pl also answer ip address being index from my earlier post

Patience - I'll get onto that once I'm happy that you're confident with
the stuff above!

------------------------------------------------------------------------------
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev

_______________________________________________
Net-snmp-users mailing list
[email protected]
Please see the following page to unsubscribe or change other options:
https://lists.sourceforge.net/lists/listinfo/net-snmp-users