Re: [Matplotlib-users] connecting data points...

[John H. <jdh...@ac...>]

>>>>> "Peter" == Peter Groszkowski <pgr...@ge...> writes:

    Peter> Hi everyone: I was wondering whether it is possible to tell
    Peter> matplotlib how/when to connect data points. Consider this
    Peter> simple script:

    Peter> from matplotlib.matlab import * figure(1) t =
    Peter> [0,1,2,3,4,5,105,106,107] s = [1,4,5,3,9,11,-5,-8,3]
    Peter> plot(t, s, antialiased=False) grid(True) show()


    Peter> There are no data points between t=5 and t=105. By default
    Peter> the points (5,11) and (105,-5) are connected, but I would
    Peter> like to tell matplotlib NOT to do so. In my case I would
    Peter> like to pass the plot function a variable telling it what
    Peter> to do. So for example would have:

    Peter> plot(t, s, max_delta=40)

    Peter> This would mean that the points are only to be connected if
    Peter> the difference between the adjacent t values is less than
    Peter> 40.  In my case this is relevant because sometimes there
    Peter> are "holes" in my data, and connecting the points makes the
    Peter> plots look very messy.

    Peter> Would anyone find something like this useful? Would it be
    Peter> difficult to implement?

Certainly not difficult, and probably useful enough to put in the
standard distro.  Eg, in a stock market trading example, you would
have lots of quotes, minute by minute, punctuated by long intervals
overnight where the market is closed.  If you set maxdelta
appropriately, you could draw connected lines only within trading
days.

Here is a sample implementation

from matplotlib.matlab import *
def segplot(x, y, fmt, maxdelta, **kwargs):
    """
    Plot x versus y, breaking the plot at any point where x[i] -
    x[i-1] > maxdelta. kwargs are passed on to plot
    """
    x = asarray(x)
    y = asarray(y)
    d = diff(x)
    lines = []
    ind = nonzero(greater(d, maxdelta))
    ind = ind+1
    if not len(ind):
        lines.extend( plot(x,y,fmt,**kwargs) ) 
    else:
        allind = [0]
        allind.extend(ind)
        allind.append(len(x))
        for i1,i2 in zip(allind[:-1], allind[1:]):
            lines.extend( plot(x[i1:i2], y[i1:i2], fmt, **kwargs) )
    return lines

t = [0,1,2,3,4,5,105,106,107,187, 200, 212, 300, 320]
s = [1,4,5,3,9,11,-5,-8,3,12, 15, 12, -1, 3]
segplot(t, s, 'b-o', 40, antialiased=False)
grid(True)
show()

I'm inclined not to make this part of plot, since plot processes a
variable number of arguments it makes it a little difficult.
Certainly doable, but I'm hesitant to put too much on plot because it
might become unwieldy.  But a new function, like segment plot, would
be easy enough to include.

Any suggestions for a name, or additional functionality?

JDH