jlp
-Interested User-
Posts: 45
Joined: Dec 10, 2007
|
Posted: Jun 22, 2011 08:33 AM
Msg. 1 of 6
Hello,
Every morning or previuous night I download 1,2,5,15,60, Daily data for about 2000 stocks and every morning I find issues where for example one of those historical data (i.e 60 min for URBN this morning) is not there for no apparent reason. This happens every day for different stocks a
Could you please tell me why this is happening and then later on the data is there.
Thanks in advance JLP
|
DTN_Steve_S
-DTN Guru-
Posts: 2093
Joined: Nov 21, 2005
|
Posted: Jun 22, 2011 08:45 AM
Msg. 2 of 6
JLP, on the request that failed, you should be getting an error message back from the feed. Can you log the actual request and result you are getting from the feed?
Also, it would be faster and easier for you to simply request 1min data and then build the rest of the bars on your end. This is the same thing that the servers do (we only store 1min data).
|
jlp
-Interested User-
Posts: 45
Joined: Dec 10, 2007
|
Posted: Jun 22, 2011 12:31 PM
Msg. 3 of 6
Hi Steve,
I don't get anything back from the feed when this happens, I have retry logic that requests up to 10 times for each interval (2,5,15,60,D) for each symbol so it retries 10 times (say URBN 60 minute this morning) and goes on the next interval or the first interval for the next symbol and then continues.
As far as downloading 1 min data, obviously that defeats the purpose of building those bars fast enough for each of those intervals.
Like I said this happens every day for one or more symbols for one or more intervals. Seems that if I request that data later on when the market opens then is ok.
I've had to code around these shortcoming but obviously is not what the feed is supposed to provide
Thanks JLP
|
DTN_Steve_S
-DTN Guru-
Posts: 2093
Joined: Nov 21, 2005
|
Posted: Jun 22, 2011 01:10 PM
Msg. 4 of 6
I'm not sure what you mean by "downloading 1 min data, obviously that defeats the purpose of building those bars fast enough for each of those intervals".
If you download 1min intervals and concatenate them into 2,5,15,60 min intervals for storage as you are parsing the data, it should be faster than retrieving the same data 5 times in a different format. Not to mention it will also be significantly less resources needed by your machine.
As far as not getting an error message is concerned, can you get me a log from IQConnect for one of your runs that shows this happening? I've sent you an email concerning how to setup the logging.
|
jlp
-Interested User-
Posts: 45
Joined: Dec 10, 2007
|
Posted: Jun 22, 2011 01:48 PM
Msg. 5 of 6
Steve,
will start logging the IQFeed starting tonight and let you know what I find.
Thanks for your help
JLP
|
jlp
-Interested User-
Posts: 45
Joined: Dec 10, 2007
|
Posted: Jun 22, 2011 01:53 PM
Msg. 6 of 6
Steve,
One more thing, I understand the suggestion about 1 minute data but the thing is that actually I only download about 10-15 minutes of 1 min data (subsequent 1 min and remaining bars are built in realltime with tick data), about 1 day of 2 min, 2 days of 5 minute, 10 days of 15 min, 20 days of 60 minute, etc so actually I don't want to have to download 1 minute data for up to 20 days since the individual requests are actually better for me. I download enough bars for each interval to have calculate MAs and other indicators that require N number of bars.
Thanks again
JLP
|
|
|
|