Re: Free RealTime Data NOW Nest ODIN Trade Tiger Google Yahoo to AmiBroker, Fcharts M
It worked. :thumb::thumb:
Damn it. In my earlier test, data from VWAP was in Descending Order :annoyed: so the first record was of 15:51:33
When I imported two records of 9:15:33 and 9:20:33, AB deleted all records from 9:15:33 and imported these two. That takes less than 1 second.
So when getting data from VWAP, we have to make sure that it is in chronological order. :annoyed:
Thanks mike for your Persistance.
Hi TB n josh,
Lots of queries But its good to have 'thinking heads', with different perspectives and individual domain expertise, to put together for general benefit.
While you guys are looking at things from a C++ / AutoIt coding perspective, with respective levels of individual domain expertise, I can only comment from amibroker's perspective.
Let me first get the basic premise of suggested format control out of the way, before I try to address other queries.
As of ami v4.7, importing data via tickmode will do the following
1. Delete the existing data in Ami from the starting point of the date timestamp of the first data point of the imported data.
i.e to say that if the time now is dd/10:45:30 in a live market and you decided to backfill from dd/09:15:00 via tickmode - Ami will delete Any existing data from (and after) dd/09:15:00 (if present in the database) till the current time.
2. And import the backfill data as a replacement.
Since you guys were getting stuck with slow execution as well as CPU hogging during tick data delete via COM calls this way was suggested, as its blazing fast and leverages the existing ami capability.
Caveat: - In live markets, during backfill in the above example, you will lose a couple of ticks from dd/10:45:30 onwards till the time it takes for the backfill of that individual ticker to complete. While it has negligible impact on price, some amount of actual transacted volume will be lost for that time period. Again, the effect is practically negligible unless a huge amount of volume decides to get transacted in those 1 - 3 seconds.
Basically you will need to couple your existing backfill code (without any coding for existing data delete) with the built in ASCII importer in ami to push data via 'tickmode'.
P.S - For starters, one can test this by importing one day, one min resolution data from a txt/csv file over the existing tick data in ami of that day via the ASCII importer, in static market conditions. Try this only for the existing last day otherwise all data ahead of that day will be deleted if any day other than the last day is chosen.
Lots of queries But its good to have 'thinking heads', with different perspectives and individual domain expertise, to put together for general benefit.
While you guys are looking at things from a C++ / AutoIt coding perspective, with respective levels of individual domain expertise, I can only comment from amibroker's perspective.
Let me first get the basic premise of suggested format control out of the way, before I try to address other queries.
As of ami v4.7, importing data via tickmode will do the following
1. Delete the existing data in Ami from the starting point of the date timestamp of the first data point of the imported data.
i.e to say that if the time now is dd/10:45:30 in a live market and you decided to backfill from dd/09:15:00 via tickmode - Ami will delete Any existing data from (and after) dd/09:15:00 (if present in the database) till the current time.
2. And import the backfill data as a replacement.
Since you guys were getting stuck with slow execution as well as CPU hogging during tick data delete via COM calls this way was suggested, as its blazing fast and leverages the existing ami capability.
Caveat: - In live markets, during backfill in the above example, you will lose a couple of ticks from dd/10:45:30 onwards till the time it takes for the backfill of that individual ticker to complete. While it has negligible impact on price, some amount of actual transacted volume will be lost for that time period. Again, the effect is practically negligible unless a huge amount of volume decides to get transacted in those 1 - 3 seconds.
Basically you will need to couple your existing backfill code (without any coding for existing data delete) with the built in ASCII importer in ami to push data via 'tickmode'.
P.S - For starters, one can test this by importing one day, one min resolution data from a txt/csv file over the existing tick data in ami of that day via the ASCII importer, in static market conditions. Try this only for the existing last day otherwise all data ahead of that day will be deleted if any day other than the last day is chosen.
Damn it. In my earlier test, data from VWAP was in Descending Order :annoyed: so the first record was of 15:51:33
When I imported two records of 9:15:33 and 9:20:33, AB deleted all records from 9:15:33 and imported these two. That takes less than 1 second.
So when getting data from VWAP, we have to make sure that it is in chronological order. :annoyed:
Thanks mike for your Persistance.
Last edited: