Re: Reg: YAJL DATA-INTO Options

[Scott Klement <rpg400-l@xxxxxxxxxxxxxxxx>]

Suren,

Why not always define the array with a 9999 size?  That seems much 
simpler and more robust.  What are you gaining by declaring it to a 
pointer and only allocating part of it?

If you did want to do it with a dynamic allocation, you could do it by 
calling YAJL's subprocedures to parse the document (i.e. don't use 
DATA-INTO, just call YAJL's routines).   This could give you the size, 
and then you could parse it a second time with DATA-INTO after you knew 
the size and have allocated the data. But this would be more 
complicated, and parsing it twice (no matter how you do it) will take 
longer than parsing it once.  So, I would simply always declare the 
array with the full 9999, I would not bother with the dynamic allocation.

-SK


On 6/19/2019 2:36 PM, Suren K wrote:
> Hi Everyone,
>
> I am using Scott's YAJL Parser using DATA-INTO opcode for Parsing the JSON
> Document. I need one clarification.
>
> Currently i am using the JSON Document which contains an array which we
> cannot limit the number of elements. So i need to define an array data
> structure with minimum 9999 elements by assuming the number of elements
> will not exceeds 9999.
>
> Instead of doing this, i am planning to use pointer based array data
> structure. Once i identified the number of elements by using the option
> "Countprefix", i can allocate the data structure based on the number of
> elements and then use the data structure in DATA-INTO to parse the full
> data.
>
> But i could see the "countprefix" option is working only when we pass the
> array Data structure together with the countprefix variable in the
> DATA-INTO Opcode.
>
> Any suggestions to how to identify the number of elements first before
> parsing the Entire Array?
>
> Regards,
> Surender K
> +1 7022091188