|
PureBytes Links
Trading Reference Links
|
Hank Lu <HLu6091187@xxxxxxx> writes:
>Your data cut-off time was about 8:20 last night, it still has a big
>gap between then and today's open... If i can fill up the gap
>between 8:20 PM last night and this morning's open, I'm sure that my
>indicators would behave just like another regular trading day.
That's because the June contract doesn't trade on globex until this
afternoon (Thursday), and I did the rollover data last night, using
what data existed up until that time on the then-active contract.
The only way to make the transition seamless would be to do the whole
process Thursday morning between 6:15am and 6:30am PST, after the
March globex closes, and before the June day session opens. Since it
takes me a couple of hours to do the job, and I'm trying to get ready
to trade at that hour of the morning, that gap is likely to be there
for a while. Of course, it can be repaired Thursday evening.
>Is it possible that you can share with me or the group on how you
>did the data conversion?
Nothing to it :)
I take the active contract tick data (SPH9 in this case) and run it
through a C program I wrote that runs on my Linux machine. The program
has arguments for what symbol I want to make, and what adjustment factor
I want to use, etc., and puts out a file in the format of TS 3.5 .dat
files. I then take that .dat file, and the 32-byte dummy name file
and put those two files in my paste-in directory and tell TS 3.5 to
Paste-In, whereupon it sees that the .idx file is missing, and kindly
builds it for me, and then offers to paste in the data from the file
I made. I delete the appropriate old tick data, and paste in the new.
That's all there is to it. Part of the time involved is rebooting
my TradeStation computer back and forth between MS Windows and Linux,
because I don't ask MS Windows to do anything other than run TS.
I've been getting the raw active-contract data from jackb (who also
helps maintain this list), from his reservoir of FutureSource data,
as a big ascii tick-data file. I could dig it out of my BMI .dat
file, but I haven't gotten around to writing that bit of code yet,
and since I wrote the FS-to-BMI converter long ago as a way to get
backup data, the path of least resistance remains this one.
Jim
|