Bug : (minor) time axis labels show "%f" instead of microseconds for years up to 1900 #3179
Labels
Milestone
Comments
|
Good spot @azjps, and thanks for making it easy to reproduce. Are you interested in having a go at implementing a fix for this? |
|
Sure :) It seems to me that these three lines in timetuple = dt.timetuple()
s1 = time.strftime(fmt, (year,) + timetuple[1:])
s2 = time.strftime(fmt, (year + 28,) + timetuple[1:])could be replaced by s1 = dt.replace(year=year).strftime(fmt)
s2 = dt.replace(year=year+28).strftime(fmt)although someone should check that setting to this year in a |
azjps
added a commit
to azjps/matplotlib
that referenced
this issue
Jul 3, 2014
azjps
added a commit
to azjps/matplotlib
that referenced
this issue
Feb 24, 2015
Add a simple test for DateFormatter. closes matplotlib#3179 Change-Id: Idff9d06cbc6dc00a3cb8dcf113983d82dbdd3fde
azjps
added a commit
to azjps/matplotlib
that referenced
this issue
Feb 25, 2015
It also fails to replace %y or %x correctly, since its strftime implementation replaces only 4-digit years. I added a boolean flag DateFormatter.replace_directives_before_1900: - If False, strftime uses the old implementation, which will not replace %f and which will replace incorrect values for %y and %x. - If True, strftime will first try a few regular expressions to replace %y/%x/%f with the appropriate datetime values. I'm not positive this covers all cases but I don't know of any cases where this fails right now. Add a simple test for DateFormatter with and without this flag. closes matplotlib#3179 Change-Id: Idff9d06cbc6dc00a3cb8dcf113983d82dbdd3fde
azjps
added a commit
to azjps/matplotlib
that referenced
this issue
May 25, 2015
It also fails to replace %y or %x correctly, since its strftime implementation replaces only 4-digit years. Instead, we now use a regular expression to replace %f with the microsecond value. (We could also be tricky and call strftime again with the original datetime object ..) We also make substitutions for both 2-digit and 4-digit years, which for example are displayed by %y, %Y, and %x. Minor point: in time.h, strftime will not use padding for %y and %Y but will use zero-padding for %x. Since this is unlikely to ever be a cause of concern and isn't documented anywhere afaik, I've used zero-padding for all three. (Certainly it's preferable to just printing the wrong year!) Add tests and (maybe excessively long?) comments. closes matplotlib#3179 Change-Id: Idff9d06cbc6dc00a3cb8dcf113983d82dbdd3fde
azjps
added a commit
to azjps/matplotlib
that referenced
this issue
May 27, 2015
It also fails to replace %y or %x correctly, since its strftime implementation replaces only 4-digit years. Instead, we now use a regular expression to replace %f with the microsecond value. (We could also be tricky and call strftime again with the original datetime object ..) We also make substitutions for both 2-digit and 4-digit years, which for example are displayed by %y, %Y, and %x. Minor point: in time.h, strftime will not use padding for %y and %Y but will use zero-padding for %x. Since this is unlikely to ever be a cause of concern and isn't documented anywhere afaik, I've used zero-padding for all three. (Certainly it's preferable to just printing the wrong year!) Add tests and (maybe excessively long?) comments. closes matplotlib#3179 Change-Id: Idff9d06cbc6dc00a3cb8dcf113983d82dbdd3fde
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
When plotting with a timestamp-based axis with datetimes of years
<= 1900, upon zooming into sub-second scales, the default axis label date formatter shows "%f" instead of microseconds.This is since
timetupledoesn't carry microsecond data. Minor in nature (was using a sentinel date of 1900 without realizing matplotlib would handle it specially) but I figured I would point it out.The text was updated successfully, but these errors were encountered: