Please note that as said in previous messages and the
Wiki page, your case of time positions (as opposed to durations)
warrants use of absolute tolerance, e.g. in milliseconds (4000 acmp).
--- Bj�rn Helgason <[EMAIL PROTECTED]> wrote:
> To make it a little more general
>
> f=: 3 : 0
> td=.(({.$bb){.y}.24 60 60 1000 #.cc) - 24 60 60 1000 #.bb
> mtd=.{.({.$bb)%~+/td
> +/|td (0.002 tcmp) mtd
> )
>
> g=: 4 : 0
> bb=: x
> cc=: y
> ($ #: (i. <./)@,) ,> f &. > i. {.$cc
> )
>
> b g c
> 20
>
> a g c
> 5
>
>
>
>
>
> 2007/3/30, Björn Helgason <[EMAIL PROTECTED]>:
> >
> >
> > So here are two verbs that do find the start for the match of a and b in
> > c
> > 5 and 20
> >
> >
> >
> > fa=: 3 : 0
> > indexc=:y
> > matchac=:lengtha{.indexc}.codc
> > timediff=:matchac - coda
> > meantimediff=:{.lengtha%~+/timediff
> > +/|timediff (0.002 tcmp) ~meantimediff
> > )
> >
> > ($ #: (i. <./)@,) ,>fa each i. lengthc
> > 5
> >
> > fb=: 3 : 0
> > indexc=:y
> > matchbc=:lengthb{.indexc}.codc
> > timediff=:matchbc - codb
> > meantimediff=:{.lengthb%~+/timediff
> > +/|timediff (0.002 tcmp) ~meantimediff
> > )
> >
> >
> > ($ #: (i. <./)@,) ,>fb each i. lengthc
> > 20
> >
> >
> > 2007/3/30, Björn Helgason <[EMAIL PROTECTED]>:
> > >
> > > This works quite nicely
> > >
> > > Here is an example from test indexc equal 5 seems to be a starting point
> > > for matching a to c matching tolerance 0.002
> > >
> > > coda=:24 60 60 1000 #.a
> > > codc=:24 60 60 1000 #.c
> > > lengtha=:$coda
> > >
> > > indexc=:0
> > >
> > > matchac=:lengtha{.indexc}.codc
> > >
> > > timediff=:matchac - coda
> > >
> > > meantimediff=:{.lengtha%~+/timediff
> > >
> > > +/|timediff (0.001 tcmp) ~meantimediff
> > > 15
> > >
> > > indexc=:5
> > >
> > > +/|timediff (0.001 tcmp) ~meantimediff
> > > 1
> > >
> > > +/|timediff (0.002 tcmp) ~meantimediff
> > > 0
> > >
> > > indexc=:10
> > >
> > > +/|timediff (0.001 tcmp) ~meantimediff
> > > 15
> > >
> > >
> > > 2007/3/29, Oleg Kobchenko <[EMAIL PROTECTED]>:
> > > >
> > > > Sorry, wrong scale direction, with increase
> > > > of radius is should become more tolerant (returning 0).
> > > >
> > > > Correct is
> > > >
> > > > tcmp=: 1 : '- ([EMAIL PROTECTED] * >&|) m*>.'
> > > >
> > > > 1134000 (0.001 tcmp) 1138000
> > > > _1
> > > > 1134000 (0.001 tcmp)~ 1138000
> > > > 1
> > > > 1134000 (0.01 tcmp) 1138000
> > > > 0
> > > >
> > > > 2 2.01 2.02 2.03 (0.01 tcmp) 2
> > > > 0 0 0 1
> > > > 2 2.01 2.02 2.03 (0.01 tcmp)&- 2
> > > > 0 0 0 _1
> > > > _1 1 (0.01 tcmp) 1 _1
> > > > _1 1
> > > >
> > > >
> > > > 2 2.001 2.01 2.1 2.5 (0.0001 tcmp) 2
> > > > 0 1 1 1 1
> > > > 2 2.001 2.01 2.1 2.5 (0.001 tcmp) 2
> > > > 0 0 1 1 1
> > > > 2 2.001 2.01 2.1 2.5 (0.01 tcmp) 2
> > > > 0 0 0 1 1
> > > > 2 2.001 2.01 2.1 2.5 (0.1 tcmp) 2
> > > > 0 0 0 0 1
> > > > 2 2.001 2.01 2.1 2.5 (0.5 tcmp) 2
> > > > 0 0 0 0 0
> > > >
> > > >
> > > > --- Bj�rn Helgason <[EMAIL PROTECTED] > wrote:
> > > >
> > > > > Right to the point as usual!
> > > > >
> > > > > 2007/3/29, Oleg Kobchenko < [EMAIL PROTECTED]>:
> > > > > >
> > > > > > 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)
> > > >
> > > >
> > > >
> > > >
> > > > ____________________________________________________________________________________
> > > >
> > > > It's here! Your new message!
> > > > Get new email alerts with the free Yahoo! Toolbar.
> > > > http://tools.search.yahoo.com/toolbar/features/mail/
> > > > ----------------------------------------------------------------------
> > > >
> > > > For information about J forums see http://www.jsoftware.com/forums.htm
> > > >
> > >
> > >
> > >
> > > --
> > > Björn Helgason, Verkfræðingur
> > > Fugl&Fiskur ehf, Ãerneyjarsund 23, Box 127
> > > 801 GrÃmsnes ,t-póst: [EMAIL PROTECTED]
> > > Skype: gosiminn, gsm: +3546985532
> > > Landslags og skrúðgarðagerð, gröfuþjónusta
> > > http://groups.google.com/group/J-Programming
> > >
> > >
> > > Tæknikunnátta höndlar hið flókna, sköpunargáfa er meistari
> > > einfaldleikans
> > >
> > > góður kennari getur stigið á tær án þess að glansinn fari af
> > > skónum
> > > /|_ .-----------------------------------.
> > > ,' .\ / | Með léttri lund verður |
> > > ,--' _,' | Dagurinn à dag |
> > > / / | Enn betri en gærdagurinn |
> > > ( -. | `-----------------------------------'
> > > | ) | (\_ _/)
> > > (`-. '--.) (='.'=)
> > > `. )----' (")_(")
> > >
> >
> >
> >
> > --
> > Björn Helgason, Verkfræðingur
> > Fugl&Fiskur ehf, Ãerneyjarsund 23, Box 127
> > 801 GrÃmsnes ,t-póst: [EMAIL PROTECTED]
> > Skype: gosiminn, gsm: +3546985532
> > Landslags og skrúðgarðagerð, gröfuþjónusta
> > http://groups.google.com/group/J-Programming
> >
> >
> > Tæknikunnátta höndlar hið flókna, sköpunargáfa er meistari
> > einfaldleikans
> >
> > góður kennari getur stigið á tær án þess að glansinn fari af skónum
> > /|_ .-----------------------------------.
> > ,' .\ / | Með léttri lund verður |
> > ,--' _,' | Dagurinn à dag |
> > / / | Enn betri en gærdagurinn |
> > ( -. | `-----------------------------------'
> > | ) | (\_ _/)
> > (`-. '--.) (='.'=)
> > `. )----' (")_(")
> >
>
>
>
> --
> Björn Helgason, Verkfræðingur
> Fugl&Fiskur ehf, Ãerneyjarsund 23, Box 127
> 801 GrÃmsnes ,t-póst: [EMAIL PROTECTED]
> Skype: gosiminn, gsm: +3546985532
> Landslags og skrúðgarðagerð, gröfuþjónusta
> http://groups.google.com/group/J-Programming
>
>
> Tæknikunnátta höndlar hið flókna, sköpunargáfa er meistari
> einfaldleikans
>
> góður kennari getur stigið á tær án þess að glansinn fari af skónum
> /|_ .-----------------------------------.
> ,' .\ / | Með léttri lund verður |
> ,--' _,' | Dagurinn à dag |
> / / | Enn betri en gærdagurinn |
> ( -. | `-----------------------------------'
> | ) | (\_ _/)
> (`-. '--.) (='.'=)
> `. )----' (")_(")
> > ----------------------------------------------------------------------
> 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
