Page 1 of 1

Data import errors

Posted: Mon Jul 26, 2010 10:02 am
by Chris_p
Hi, I've been trying to import some data that I downloaded from Dukascopy, but appear to be getting some errors. The data is in 1 minute intervals, but only shows up as daily bars when imported. A sample of the data is attached.

Thanks,
Chris

Posted: Mon Jul 26, 2010 10:04 am
by Chris_p
File didn't attach the first time. Here it is...

Posted: Mon Jul 26, 2010 10:05 am
by Chris_p
Date Time Open High Low Close Volume
2009.07.01 00:00:00 1.40395 1.40425 1.4039 1.40425 102.4
2009.07.01 00:01:00 1.40425 1.4043 1.40405 1.4041 182.2
2009.07.01 00:02:00 1.40415 1.40435 1.40415 1.40425 175.8
2009.07.01 00:03:00 1.40425 1.40435 1.40415 1.40425 68.2
2009.07.01 00:04:00 1.40425 1.4043 1.40405 1.40405 80.4
2009.07.01 00:05:00 1.40405 1.4042 1.4038 1.4038 151.4
2009.07.01 00:06:00 1.4038 1.40385 1.403 1.403 189.6
2009.07.01 00:07:00 1.403 1.40375 1.403 1.4033 334.2
2009.07.01 00:08:00 1.4033 1.4033 1.40305 1.4032 103.4
2009.07.01 00:09:00 1.4034 1.40355 1.40325 1.40335 122.8
2009.07.01 00:10:00 1.40345 1.40355 1.40325 1.40325 243
2009.07.01 00:11:00 1.40325 1.4035 1.40325 1.40345 184.6
2009.07.01 00:12:00 1.40345 1.4036 1.40335 1.4036 127
2009.07.01 00:13:00 1.40355 1.40375 1.4035 1.40375 267.4
2009.07.01 00:14:00 1.40375 1.40385 1.40365 1.4037 255.6
2009.07.01 00:15:00 1.4038 1.4038 1.40365 1.4037 77.6
2009.07.01 00:16:00 1.4037 1.4037 1.40335 1.4034 142.8
2009.07.01 00:17:00 1.4034 1.40345 1.40335 1.40335 117.5
2009.07.01 00:18:00 1.40335 1.40355 1.4031 1.4031 202.3
2009.07.01 00:19:00 1.40315 1.4035 1.40315 1.40335 262.4
2009.07.01 00:20:00 1.40335 1.40365 1.40335 1.4036 237.1

Posted: Tue Jul 27, 2010 4:50 am
by FT Support
Sorry but Forex Tester does not support such data format, i'll try to find a programmer who will be able to create data converter to translate data into FT format

Posted: Tue Jul 27, 2010 6:16 am
by Chris_p
FT Support wrote:Sorry but Forex Tester does not support such data format, i'll try to find a programmer who will be able to create data converter to translate data into FT format


Thanks.

So what's wrong with the data format? If I know where the errors are, I can maybe correct them myself. Should the file be converted to a txt file?

Posted: Wed Jul 28, 2010 8:05 am
by FT Support
Please see incompatibilities of these data below:

1) Date format should be one of the followng: YYYYMMDD, DDMMYYYY, MMDDYYYY. YYMMDD, DDMMYY, MMDDYY

2) Time format should be HHMMSS

Posted: Sat Jul 31, 2010 1:53 am
by Chris_p
FT Support wrote:Please see incompatibilities of these data below:

1) Date format should be one of the followng: YYYYMMDD, DDMMYYYY, MMDDYYYY. YYMMDD, DDMMYY, MMDDYY

2) Time format should be HHMMSS


Ok, thanks. I'll see if I can change it and get it to work.

Importing Dukascopy 1 Minute Data

Posted: Sat Jul 02, 2011 10:17 pm
by SquareMeal
I've developed a full-bodied system in FT which I want to trade live. I opened a Dukascopy account and converted my "execution" module to JForex -- but the rest of my system (multiple modules for optimizing) is in FT. Things have worked out pretty well, but there are some differences between execuction in FT and the JForex platform. So, I converted Dukascopy data to load into FT to eliminate that variable from the problem.

I think the data conversion went well (a java program I wrote), but when I run my system and compare 30 minute bar charts between FT and JForex, there are some significant differences. Also, focusing on a single bar (2007-09-07 02:30), the one produced in FT doesn't seem to match the OHLC within that 30 minute period... but the JForex bar seems to reflect the data.

I'm attaching two annotated screen shots. Can you please tell me if I have a data conversion problem? Thank you.

Posted: Mon Jul 04, 2011 1:24 pm
by FT Support
Hello,

Pleae check if you have same M1 bars in JForex and FT in History Mode than check if you have same M1 bars in Testing Mode and let us know.

M1 data conversion

Posted: Wed Jul 20, 2011 8:47 pm
by SquareMeal
I will still try your suggeston regarding 1min data. But for now, I should say that I changed my conversion program to round volume to an integer (got rid of the decimal point) and the results are 98% better. There are still some wicks on the candles that are different, but the candle bodies now look the same without all of the gaps I was getting.