Maybe you need to convert the timestamps to
integers and use a scaled tolerant comparison:

   tcmp=: 1 : '(m*>.) (>&| * [EMAIL PROTECTED]) -'
   1134000 (0.01 tcmp)  1138000
_1
   1134000 (0.001 tcmp)  1138000
0
   1134000 (0.01 tcmp)~  1138000
1

This is by extended definition of ''tolerance''.

http://www.google.com/search?q=tolerance&sitesearch=jsoftware.com



--- Bj�rn Helgason <[EMAIL PROTECTED]> wrote:

> Lets say that the problem is to do a fuzzy match
> Not exact match.
> As you see the time difference is sometimes 18 56, 18 55, 18 57 so a good
> match is something like 18 56 +/- 1  or 2 seconds
> It would not be any problem if the difference is always exactly 18 56
> The problem is also finding what the time difference is and in what
> direction.
> Time difference by some measurements outside the few second span would not
> be good.
> 
> So getting a 0/1 answer would rather be something like an interval of a few
> seconds for the timedifferences in each case
> 
> This problem is fairly trivial to do manually.
> 
> Finding the probe by looking at the image was also trivial manually.
> Seeing if the probe was there and if it was in the center or in the
> periphery.
> Doing it by a program seemed to me like an impossibility a month ago.
> That part is the easy part now.
> 
> Doing this fuzzy matching is probably easy as well even if it is not crystal
> clear to me right now.
> The differences in time are probably triggered by that the probe is
> triggered by receiving exposure and it seems to vary a second or two.
> Also the clock in each equipment is not in sync.
> 
> The correct match could be within quite a span if the number is coded
> 
>      24 60 60 1000 #.0 18 54 0
> 1134000
>     24 60 60 1000 #.0 18 58 0
> 1138000
> 
> 2007/3/29, Devon McCormick <[EMAIL PROTECTED]>:
> >
> > So, taking your guess here of the time offset of 18 minutes, 56 seconds= 0
> > 18 56 0, I can
> > adjust c:
> >
> >    3{.adjc=. 24 60 60 1000 #: (24 60 60 1000 #.c)-24 60 60 1000 #. 0 18 56
> > 0
> > 10 57 51 450
> > 10 57 59 150
> > 10 58 6 950
> >
> > and come up with a boolean to interpolate it into a,b:
> >
> >    (#adjc)>gv=. /: adjc,a,b
> > 1 1 1 1 1 0 1 1 0 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 1 0
> > 0
> > 1 0 1 1 0 0 1 0 1 0 1 0 1 1 0 0 1 0 1 0 1 1 0 1 0 0 1
> >
> > where the 1s mark successive elements of c and the 0s mark successive
> > elements of a,b.  Is this any help?
> >
> >
> > On 3/29/07, Bj�rn Helgason <[EMAIL PROTECTED]> wrote:
> > >
> > > All of c come from a CT or similar equipment.
> > > a and b come from a instrument that are put into the CT.
> > >
> > > The CT gives exposure and in c come the times - and lots of other
> > > information when the exposure is made.
> > >
> > > a and b come from the instrument - a probe in a phantom - measure doses
> > > and
> > > such and also register time.
> > >
> > > I have been making program to find the place where the probe was on the
> > > image.
> > >
> > > I have gathered the information from different files I get.
> > >
> > > I may not be getting the correct files, files may be missing.
> > >
> > > I know a,b and c came from the same session.
> > >
> > > The person did some exposures before the probe was connected so some
> > > values
> > > are missing.
> > >
> > > It is quite obvious that the clock in the probe (a and b) and in the CT
> > > (c)
> > > are not in sync.
> > > The time difference is ca 18 minutes and 56 seconds as you can see doing
> > > the
> > > calculations.
> > >
> > > I have a gap at the beginning introducing d to fill the gap.
> > > The gap could have been in the middle.
> > > I have no way of knowing in advance.
> > >
> > > I am trying to make a reasonable guess which times fit so I can match
> > the
> > > exposure doses gathered from the probe (a and b) and not having to do it
> > > manually.
> > >
> > > The time difference is not exactly 18 minutes and 56 seconds between all
> > > corresponding lines in c and the lines in a and b. I do not know in
> > > advance
> > > if I do have a time difference and I do not know which way the
> > difference
> > > is
> > > going to be.
> > >
> > > It is fair to assume all the times in c to be sorted.
> > > It is also fair to assume the the individual files a and b are sorted.
> > >
> > > I have to find where lines in a matches lines in c
> > > I have to find where lines in b matches lines in c
> > >
> > > The time difference in all cases should be roughly the same.
> > >
> > > I guess if I get a file - lets call it e - I would have to try to be
> > able
> > > to
> > > make an educated guess that I could not find any match and it would be
> > > from
> > > some other test.
> > >
> > > Hope that explains the task at hand fairly well.
> > >
> > > Up to now I have done this part manually but would like to be able to
> > see
> > > if
> > > I could get this to work.
> > > I have never tried to match something like this with a program before.
> > > Most of the time there is an exact match or there is not.
> > > Not such a fuzzy match of matching roughly.
> > >
> > > Obviously I have to match the dates as well and both the formats of the
> > > date
> > > and the time were different.
> > > I also added 000 in both a and b
> > > If I am not lucky the test person could be sending in files from
> > different
> > > tests.
> > > I am working on what requirements the test persons are supposed to have.
> > > Try to exclude complete morons somehow politely.
> > >
> > > 2007/3/29, Devon McCormick <[EMAIL PROTECTED]>:
> > > >
> > > > Gosi - I don't understand what you're trying to do here.
> > > > Are these timestamps that need to be matched?  If so, it looks like
> > > > all of "c" comes after everything in "a" and "b".
> > > >
> > > >    gv=. /:c,a,b
> > > >    gv<#c
> > > > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1
> > 1
> > > 1
> > > > 1
> > > > 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
> > > >
> > > > Also, is there any reason to complicate the problem by having a
> > separate
> > > > "a"
> > > > and "b"?
> > > > Aren't you dealing with ab=. a,b versus c?
> > > >
> > > > Maybe I need either more context or a simpler version of this.
> > > >
> > > > On 3/29/07, Bj�rn Helgason <[EMAIL PROTECTED]> wrote:
> > > > >
> > > > > Does anyone have a good way of matching times?
> > > > >
> > > > > Here I have an example they happen to be somewhat in order.
> > > > > c contains 35 lines that need to be matched with a and b.
> > > > > There is an excess of 5 lines in c compared to a+b so d contains 5
> > > > > dummies.
> > > > >
> > > > > The times need not be sorted like this and they may be missing in
> > > either
> > > > > direction.
> > > > >
> > > > > The task is to get a best match which gives the least time
> > difference
> > > > and
> > > > > get the indexes for c that give the best match for a+b
> > > > >
> > > > > There are a lot of fields in each line but only connection between
> > > them
> > > > is
> > > > > the time field.
> > > > > In this example the time difference between the equipment is roughly
> > > 19
> > > > > minutes
> > > > >
> > > > >
> > > > >
> > > > > NB. -----------------head
> > > > >
> > > > > a=:".;._2 (0 : 0)
> > > > > 11 00 29 000
> > > > > 11 00 38 000
> > > > > 11 00 44 000
> > > > > 11 00 52 000
> > > > > 11 01 00 000
> > > > > 11 03 20 000
> > > > > 11 03 28 000
> > > > > 11 03 35 000
> > > > > 11 03 43 000
> > > > > 11 03 51 000
> > > > > 11 05 09 000
> > > > > 11 05 17 000
> > > > > 11 05 25 000
> > > > > 11 05 32 000
> > > > > 11 05 40 000
> > > > > )
> > > > >
> > > > > NB. -------------------- abdomen
> > > > >
> > > > > b=:".;._2 (0 : 0)
> > > > > 11 11 04 000
> > > > > 11 11 11 000
> > > > > 11 11 19 000
> > > > > 11 11 27 000
> > > > > 11 11 34 000
> > > > > 11 12 46 000
> > > > > 11 12 54 000
> > > > > 11 13 02 000
> > > > > 11 13 10 000
> > > > > 11 13 17 000
> > > > > 11 14 36 000
> > > > > 11 14 44 000
> > > > > 11 14 52 000
> > > > > 11 15 00 000
> > > > > 11 15 07 000
> > > > > )
> > > > >
> > > > > NB. ----------------- dcm
> > > > >
> > > > > c=:".;._2 (0 : 0)
> > > > > 11 16 47 450
> > > > > 11 16 55 150
> > > > > 11 17 02 950
> > > > > 11 17 10 600
> > > > > 11 17 18 300
> > > > > 11 19 25 350
> > > > > 11 19 33 200
> > > > > 11 19 41 000
> > > > > 11 19 48 700
> > > > > 11 19 56 450
> > > > > 11 22 16 250
> > > > > 11 22 24 050
> > > > > 11 22 31 750
> > > > > 11 22 39 500
> > > > > 11 22 47 200
> > > > > 11 24 05 650
> > > > > 11 24 13 450
> > > > > 11 24 21 150
> > > > > 11 24 28 850
> > > > > 11 24 36 600
> > > > > 11 29 59 550
> > > > > 11 30 07 350
> > > > > 11 30 15 050
> > > > > 11 30 22 750
> > > > > 11 30 30 450
> > > > > 11 31 42 600
> > > > > 11 31 50 400
> > > > > 11 31 58 100
> > > > > 11 32 05 850
> > > > > 11 32 13 550
> > > > > 11 33 32 400
> > > > > 11 33 40 200
> > > > > 11 33 47 900
> > > > > 11 33 55 600
> > > > > 11 34 03 300
> > > > > )
> > > > >
> > > > > NB. ------------------ auka
> > > > > d=:".;._2 (0 : 0)
> > > > > 0 0 0 0
> > > > > 0 0 0 0
> > > > > 0 0 0 0
> > > > > 0 0 0 0
> > > > > 0 0 0 0
> > > > > )
> > > > >
> > > > > NB. ------------------- example calculations - the last one gives
> > the
> > > > > correct lines hooked together
> > > > > c-a,b,d
> > > > > 24 60 60 1000 #:((24 60 60 1000 #.c)-24 60 60 1000 #.a,b,d)
> > > > >
> > > > > c-~a,b,d
> > > > > 24 60 60 1000 #:((24 60 60 1000 #.c)-~24 60 60 1000 #.a,b,d)
> > > > >
> > > > > c-d,a,b
> > > > > 24 60 60 1000 #:((24 60 60 1000 #.c)-24 60 60 1000 #.d,a,b)
> > > > >
> > > > >
> > > > >
> > > > >
> > > > > --
> > > > > Bj�rn Helgason, Verkfr��ingur
> > > > > Fugl&Fiskur ehf, �erneyjarsund 23, Box 127
> > > > > 801 Gr�msnes ,t-p�st: [EMAIL PROTECTED]
> > > > > Skype: gosiminn, gsm: +3546985532
> > > > > Landslags og skr��gar�ager�, gr�fu�j�nusta
> > > > > http://groups.google.com/group/J-Programming
> > > > >
> > > > >
> > > > > T�knikunn�tta h�ndlar hi� fl�kna, sk�punarg�fa er 
> > > > > meistari
> > > > einfaldleikans
> > > > >
> > > > > g��ur kennari getur stigi� � t�r �n �ess a� glansinn 
> > > > > fari af sk�num
> > > > >           /|_      .-----------------------------------.
> > > > >          ,'  .\  /  | Me� l�ttri lund ver�ur        |
> > > > >      ,--'    _,'   | Dagurinn � dag                     |
> > > > >     /       /       | Enn betri en g�rdagurinn  |
> > > > >    (   -.  |        `-----------------------------------'
> > > > >    |     ) |        (\_ _/)
> > > > >   (`-.  '--.)       (='.'=)
> > > > >    `. )----'        (")_(")
> > > > >
> > > > >
> > ----------------------------------------------------------------------
> > > > > For information about J forums see
> > http://www.jsoftware.com/forums.htm
> > > > >
> > > >
> > > >
> > > >
> > > > --
> > > > Devon McCormick, CFA
> > > > ^me^ at acm.
> > > > org is my
> > > > preferred e-mail
> > > > ----------------------------------------------------------------------
> > > > For information about J forums see http://www.jsoftware.com/forums.htm
> > > >
> > >
> > >
> > >
> > > --
> > > Bj�rn Helgason, Verkfr��ingur
> > > Fugl&Fiskur ehf, �erneyjarsund 23, Box 127
> > > 801 Gr�msnes ,t-p�st: [EMAIL PROTECTED]
> > > Skype: gosiminn, gsm: +3546985532
> > > Landslags og skr��gar�ager�, gr�fu�j�nusta
> > > http://groups.google.com/group/J-Programming
> > >
> > >
> > > T�knikunn�tta h�ndlar hi� fl�kna, sk�punarg�fa er meistari
> > einfaldleikans
> > >
> > > g��ur kennari getur stigi� � t�r �n �ess a� glansinn fari 
> > > af sk�num
> > >           /|_      .-----------------------------------.
> > >          ,'  .\  /  | Me� l�ttri lund ver�ur        |
> > >      ,--'    _,'   | Dagurinn � dag                     |
> > >     /       /       | Enn betri en g�rdagurinn  |
> > >    (   -.  |        `-----------------------------------'
> > >    |     ) |        (\_ _/)
> > >   (`-.  '--.)       (='.'=)
> > >    `. )----'        (")_(")
> > >
> > > ----------------------------------------------------------------------
> > > For information about J forums see http://www.jsoftware.com/forums.htm
> > >
> >
> >
> >
> > --
> > Devon McCormick, CFA
> > ^me^ at acm.
> > org is my
> > preferred e-mail
> > ----------------------------------------------------------------------
> > For information about J forums see http://www.jsoftware.com/forums.htm
> >
> 
> 
> 
> -- 
> Bj�rn Helgason, Verkfr��ingur
> Fugl&Fiskur ehf, �erneyjarsund 23, Box 127
> 801 Gr�msnes ,t-p�st: [EMAIL PROTECTED]
> Skype: gosiminn, gsm: +3546985532
> Landslags og skr��gar�ager�, gr�fu�j�nusta
> http://groups.google.com/group/J-Programming
> 
> 
> T�knikunn�tta h�ndlar hi� fl�kna, sk�punarg�fa er meistari 
> einfaldleikans
> 
> g��ur kennari getur stigi� � t�r �n �ess a� glansinn fari af 
> sk�num
>           /|_      .-----------------------------------.
>          ,'  .\  /  | Me� l�ttri lund ver�ur        |
>      ,--'    _,'   | Dagurinn � dag                     |
>     /       /       | Enn betri en g�rdagurinn  |
>    (   -.  |        `-----------------------------------'
>    |     ) |        (\_ _/)
>   (`-.  '--.)       (='.'=)
>    `. )----'        (")_(")
> > ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm



 
____________________________________________________________________________________
Expecting? Get great news right away with email Auto-Check. 
Try the Yahoo! Mail Beta.
http://advision.webevents.yahoo.com/mailbeta/newmail_tools.html 
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to