Datetime64 fixes #816
Datetime64 fixes #816
Conversation
philippjfr
added
type: bug
| @@ -61,6 +63,14 @@ def get_data(self, element, ranges, style): | |||
| ys = element.dimension_values(1) | |||
| return (xs, ys), style, {} | |||
|
|
|||
| def init_artists(self, ax, plot_args, plot_kwargs): | |||
| xs, ys = plot_args | |||
| if xs.dtype.kind == 'M': | |||
There was a problem hiding this comment.
Might be worth adding a comment mentioning what the 'M' code means. I assume it is the numpy datetime64 type?
|
Looks good. Ready to merge? |
|
Think there might be a few more places where datetime64 support can be improved. Will investigate a bit more. |
|
I've now overhauled the PR again. The main changes here are that you can now register datetime formatters for datetime64 and datetime types, which means you can now do this: %%opts Curve [xrotation=15 xticks=5]
import numpy as np
from datetime import date
import pandas as pd
start = date(2012, 1, 15)
end = date(2012, 1, 30)
hv.Dimension.type_formatters[np.datetime64] = '%m/%d'
hv.Curve((pd.date_range(start, end, freq='D'), np.random.rand(16)))
Additionally I fixed various bugs in xarray date range handling and made formatting of grid tick labels consistent with everything else. Ready to merge now, but there's a definite possibility tests will break due to Grid tick formatting changes. |
|
Ready to review, updated test data will pass shortly. |
| @@ -262,9 +262,6 @@ class GridPlot(CompositePlot): | |||
| show_legend = param.Boolean(default=False, doc=""" | |||
| Legends add to much clutter in a grid and are disabled by default.""") | |||
|
|
|||
| tick_format = param.String(default="%.2f", doc=""" | |||
There was a problem hiding this comment.
So this parameter is now deprecated?
There was a problem hiding this comment.
Hmm, yes, we have a system for tick formatting using Dimension.value_format and Dimension.type_formatters, so this is redundant. You think we need to go through a deprecation cycle for this?
There was a problem hiding this comment.
It will at least need to be mentioned in the CHANGELOG. Other than that I think it is fine.
|
Looks good. I've made two comments and once you have addressed those, I am happy to merge. |
|
This pull request has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs. |
Matplotlib does not currently support plotting numpy datetime64 types directly. This PR adds a compatibility layer ensuring these types are converted appropriately for plotting in matplotlib.