Paul,
You were absolutely right. I had downloaded older data from Yahoo!, then filled in more recent data from a paid service (since it has a limited backfill history). The period in which I have mixed data is the period for each stock that is messed up. Thanks so much! Cheers, Ed From: [email protected] [mailto:[EMAIL PROTECTED] On Behalf Of paultsho Sent: Thursday, March 27, 2008 7:56 AM To: [email protected] Subject: [amibroker] Re: Problems with Foreign() Ed, I now know what your problem is. Google has data holes, not the type you would normally preceived as data holes but nonetheless they are. from 3/26 on, you no longer have daily bars, so AB displays the compressed intraday bars(thats why you have time stamps), but since you dont have intraday bars of ~AdvIss to compress. foreign of course displays zero volume. Your problem can be fixed by importing EOD data into your mixed database, but a better solution would be to separate intraday and EOD data. Addtocomposite is very strict on bar alignment. A separate EOD database solves a lot of problrms --- In [email protected] <mailto:amibroker%40yahoogroups.com> , "Ed Cottrell" <[EMAIL PROTECTED]> wrote: > > Hi Paul, > > > > I tried creating an exploration as you said. The output for V is always > correct for every symbol viewed, including ~AdvIss. The output of > Foreign("~AdvIss", "V"), however, varies - when running the exploration on > ~AdvIss, it's correct, but for other symbols, it has holes. Here's my > exploration code: > > Filter = 1; > > AddColumn(V, "Volume"); > > AddColumn(Foreign("~AdvIss","V"), "AdvIss"); > > > > The following is the output when I run the exploration on GOOG over the > window 7/20/07 to 7/31/07 (the data hole when exploring on GOOG runs from > 7/26/07 to 12/31/07): > > Ticker Date/Time Volume AdvIss > > GOOG 7/20/2007 35547612.00 162.00 > > GOOG 7/23/2007 12705503.00 665.00 > > GOOG 7/24/2007 11313526.00 89.00 > > GOOG 7/25/2007 11101426.00 695.00 > > GOOG 7/26/2007 9:30:00 AM 6894330.00 0.00 > > GOOG 7/27/2007 9:30:00 AM 5509031.00 0.00 > > GOOG 7/30/2007 9:30:00 AM 3967248.00 0.00 > > GOOG 7/31/2007 9:30:00 AM 4270981.00 0.00 > > > > > > The same time frame over ~AdvIss gives: > > Ticker Date/Time Volume AdvIss > > ~AdvIss 7/20/2007 162.00 162.00 > > ~AdvIss 7/23/2007 665.00 665.00 > > ~AdvIss 7/24/2007 89.00 89.00 > > ~AdvIss 7/25/2007 695.00 695.00 > > ~AdvIss 7/26/2007 100.00 100.00 > > ~AdvIss 7/27/2007 220.00 220.00 > > ~AdvIss 7/30/2007 1014.00 1014.00 > > ~AdvIss 7/31/2007 294.00 294.00 > > > > > > Any thoughts on where to go from here? Thanks! > > > > -Ed > > > > From: [email protected] <mailto:amibroker%40yahoogroups.com> [mailto:[email protected] <mailto:amibroker%40yahoogroups.com> ] On Behalf > Of Paul Ho > Sent: Tuesday, March 25, 2008 5:57 AM > To: [email protected] <mailto:amibroker%40yahoogroups.com> > Subject: RE: [amibroker] Problems with Foreign() > > > > There are many possible reasons for this to happen, and it is impossible to > guess what it is. What I can suggest is a way to find out what causes the > problem yourself. > > Create an exploration usng addcolumn to list out the value of V, do that on > various symols that you would view from. including the ticker ~AdvIss > itself, and find out what sort of volume you would get. and take it from > there. > > > > _____ > > From: [email protected] <mailto:amibroker%40yahoogroups.com> [mailto:[email protected] <mailto:amibroker%40yahoogroups.com> ] On Behalf > Of Ed Cottrell > Sent: Tuesday, 25 March 2008 5:13 AM > To: 'AmiBroker Group' > Subject: [amibroker] Problems with Foreign() > > Hello all, > > I am using Foreign() to calculate STIX (an advance/decline stat). I have > already used AddToComposite to count advances and declines over a certain > group of stocks and store that info in the volume for ~AdvIss and ~DecIss. > > Here's my problem: if I plot one of those composites using > Foreign("~AdvIss", "V"), the output depends on which stock I am viewing in > my main chart. For some stocks, I get the actual volume number stored in > ~AdvIss; for the rest, I get zero. > > What am I doing wrong here? > > Thanks, > > Ed >
