 Posted 2/23/2007 18:34:00
|
|
|
|
| Ok Thanks, I will try it out.
|
|
|
Posted 2/23/2007 18:55:39
|
|
|
|
| Great! I'll message you when there's something available. Things are a little broken at the moment.
|
|
|
Posted 2/23/2007 19:30:23
|
|
|
|
| Thanks Bill. You being able to run it, made me curious; So I downloaded the latest beta and tested it. And what you know, it works fine. Now it brings two possibilities that I will try to narrow down. One is that: As I had a long range for the data on the previous test and installation; It is possible that the range of data was missing bars for either of the two symbols. But I should have got warrings for that. Or, the range and data size is troubling something in RE. Another is that: Something could have been missing in my version, as I was testing with an interim one also. However, I suspect the former.
|
|
|
Posted 2/23/2007 20:11:37
|
|
|
|
| Yep, the problem was the range of the data. The original data set went back to 1/1/1980 and that fetched; ^GSPC: 6850 bars and ^XAU: 5847. So these two didn't align in data size. We can short the testing range, or check for the size of the bars for symbols and if not equal, shift the start of data filling by the difference.
|
|
|
Posted 2/23/2007 21:17:55
|
|
|
|
| OK, that does make sense since my system was starting at 1/1/2004. Did you at least get the squawk that there were missing bars? Siasb (2/23/2007)
Yep, the problem was the range of the data. The original data set went back to 1/1/1980 and that fetched; ^GSPC: 6850 bars and ^XAU: 5847. So these two didn't align in data size. We can short the testing range, or check for the size of the bars for symbols and if not equal, shift the start of data filling by the difference.
|
|
|
Posted 2/24/2007 14:56:32
|
|
|
|
| No I didn't get that. But since I am using series directly, that might be the reason for not getting them.
|
|
|
Posted 2/24/2007 19:38:00
|
|
|
|
The missing bar squawk only comes up if there are gaps in your bar data. If the start date for different symbols differ, this is not considered a gap. But, you will get empty bars for a symbol when simulating a date before the start of its data.
To work around this, you should add double.NaN to your "SPX" series whenever your spxBar is empty. That will make the data line up.
We may change how this works so that you do not have to add double.NaN to your series manually. Of course if we do this, you would probably have to change your code back to the way it is currently 
Thanks,
Daniel
|
|
|
|