Divide Large Data Blob?
bobsneidar at iotecdigital.com
Mon May 16 18:44:44 EDT 2022
So this has got me thinking. Apparently what I am calling Divide and Conquer is really called a binary sort. I have looked up on the interwebs to calculate the maximum number of iterations for a given number of values, but it seems that all the formulas offered up use functions for C. I am trying to figure out what a basic math formula for this is, given n values.
> On May 16, 2022, at 15:23 , Bob Sneidar via use-livecode <use-livecode at lists.runrev.com> wrote:
> A maximum of 7 recursions are necessary to isolate a single instance of 100 possible values. 1000 requires a maximum of 10. 10000 values requires 14. The idea is that for every factor of 10, you need roughly 3 more recursions. This of course assumes the data is sorted, which in your case is sorted into 3 containers. If you know the limits of how many lines can be garbage, and how many can be valid data, you narrow your scope significantly.
> Livecode is pretty damn quick at parsing this kind of data. If there are consistent delimiters (in this case a line break) then even 20 or 30 recursions is child's play.
> Bob S
>> On May 16, 2022, at 15:00 , Bob Sneidar via use-livecode <use-livecode at lists.runrev.com> wrote:
>> Do you know exactly which lines you need to toss, or do you need to searc the data to find out where the beginning and end of the useful data is?
>> If the former, then just put line x to y of your data into a new variable. If the latter, then a divide and conquer approach might be the answer. Get the line 30% in, test for valid, get the line 40% in, test, then 35% then 32.5% or 37.5% depending on your test.
>> You may only have to do this a dozen or so times to find the exact line where your valid data begins.
>> The other way of course is to get it all into a SQL database (how did you all know I was going to say that??) The downside is that you have to iterate through all your data once. The upside is a good one liner query statement may be all you need to process your data. And if you need to make multiple passes at your data, all the better.
>> Bob S
>>> On May 16, 2022, at 10:46 , Rick Harrison via use-livecode <use-livecode at lists.runrev.com> wrote:
>>> I have a large chunk of data that I want to
>>> search as quickly as possible.
>>> Unfortunately the part I want to search is the
>>> middle third of the data. The other thirds at
>>> the beginning and at the end are just junk and
>>> slow down my search so I want to get rid of them.
>>> I don’t want to search line by line as that
>>> takes way too long.
>>> There’s no unique character dividing any
>>> of these data regions.
>>> What’s the best way to do this?
>>> Thanks in advance!
>>> use-livecode mailing list
>>> use-livecode at lists.runrev.com
>>> Please visit this url to subscribe, unsubscribe and manage your subscription preferences:
>> use-livecode mailing list
>> use-livecode at lists.runrev.com
>> Please visit this url to subscribe, unsubscribe and manage your subscription preferences:
> use-livecode mailing list
> use-livecode at lists.runrev.com
> Please visit this url to subscribe, unsubscribe and manage your subscription preferences:
More information about the use-livecode