From: Stephen W. <ste...@cs...> - 2004-08-05 17:49:03
|
On Thu, 2004-08-05 at 10:00, John Hunter wrote: > >>>>> "Jin-chung" =3D=3D Jin-chung Hsu <hs...@st...> writes: >=20 > Jin-chung> When I do the bar(left, height) in matlab the first > Jin-chung> time, e.g.: > >>>> bar([4,5,6,7,8],[9,8,7,6,5]) >=20 > This was a bug in the way I update the datalim in axes.py, which is > currently a bit too fragile for my tastes. Below I'll include the > modified matplotlib.axes.Axes.bar function. Let me know if it passes > all your tests... I was interested in this report, and sorry if I butted in. John's patch fixes Jin-chung's problem with poor x-axis scaling here. I'm using CVS, not 0.60.2, and didn't see the "funny Y tick labels" he reported. However, in either case, the bars' left edge is at the x coordinate rather than being centered on it. I know this is the documented behavior, but Matlab centers the bars on the given X coordinates. How do people feel about changing matplotlib to match the Matlab behavior?=20 I've changed my copy locally. --=20 Stephen Walton <ste...@cs...> Dept. of Physics & Astronomy, Cal State Northridge |