[Matplotlib-users] First impression from a new user

[Marquardt, Colin]

Title: First impression from a new user

[Repost, sorry if you get this twice.]

Hi guys,

in my quest for a better gnuplot replacement, I came across matplotlib
yesterday.

I really like it, big thanks to the developers.

It was pretty easy to port over a gnuplot command file.  There are
two things that could have been easier still though:

* My data file consists of a date (in ISO format) and integers. Parsing
  the date was a bit of work. I understand that python's datetime doesn't
  provide any parsing of dates - but maybe matplotlib should have some
  functions for that then?  FWIW, here is how gnuplot does this (adapted
  from http://t16web.lanl.gov/Kawano/gnuplot/datetime-e.html):

     set xdata time
     set timefmt "%Y-%m-%d"
     set format x "%Y-%m"

  In my matplotlib script, I'm now doing:

     mydate = [int(elem) for elem in vals_line[0].split('-')]
     datenums.append(date2num(datetime.date(mydate[0], mydate[1], mydate[2])))
     [...]
     plotline = plot_date(datenums, data)

* gnuplot has a plot style "steps"
  (http://t16web.lanl.gov/Kawano/gnuplot/intro/style-e.html) - I
  could only fake that with with extra "dummy" points in matplotlib
  as I understand it. It would be nice if matplotlib could do this
  for me.

Cheers,
  Colin