How Timelines Unify in subplot

Question

peter huang 2022-9-24

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1811230-how-timelines-unify-in-subplot

编辑： dpb 2022-9-25

I want to unify the time of the xtick in the subplot, but the time length of the data is different I adjusted through the axis but the graphs that have not yet reached 2021 cannot appear in 2021

test.m code:

clear all ; clc ; clf

load compound_tide_1516_data_wl_3.mat

load compound_tide_1516_data_wl_5.mat

subplot(211)

plot(time_point_1516_wl_3 ,moving_M2_ht_wl_3,'r--*' );hold on

plot(time_point_1516_wl_3 ,moving_S2_ht_wl_3,'b--*' );hold on

plot(time_point_1516_wl_3 ,moving_K1_ht_wl_3,'c--*' );hold on

plot(time_point_1516_wl_3 ,moving_O1_ht_wl_3,'m--*' );hold on

datetick('x','yyyy')

axis([time_point_1516_wl_3(1) time_point_1516_wl_3(end)+200 0.0 0.30])

text1 = text(time_point_1516_wl_3(1)+100,0.27,'WL:3 yr')

% axis([time_point_1516_wl_3(1) time_point_1516_wl_3(end)+600 0.0 0.30])

set(text1,'FontSize',15)

set(gca,'FontSize',15)

subplot(212)

plot(time_point_1516_wl_5 ,moving_SA_ht_wl_5,'k--*' );hold on

plot(time_point_1516_wl_5 ,moving_M2_ht_wl_5,'r--*' );hold on

plot(time_point_1516_wl_5 ,moving_S2_ht_wl_5,'b--*' );hold on

plot(time_point_1516_wl_5 ,moving_K1_ht_wl_5,'c--*' );hold on

plot(time_point_1516_wl_5 ,moving_O1_ht_wl_5,'m--*' );hold on

% datetick('x',28)

% axis([time_point_1516_wl_3(1) time_point_1516_wl_3(end) 0.0 0.30])

datetick('x','yyyy')

text2 = text(time_point_1516_wl_3(1)+100,0.27,'WL:5 yr')

set(text2,'FontSize',15)

set(gca,'FontSize',15)

3 个评论
显示 1更早的评论隐藏 1更早的评论

peter huang 2022-9-25

What I mean is that I want to display 2008-2022 in both the subplot(211) and subplot(212) xticks at the beginning and end, but the data of subplot(212) only reaches 2020, so 2022 has not been displayed on xtikc and can you elaborate on why I should use datetime instead

dpb 2022-9-25

编辑：dpb 2022-9-25

"...I want to display 2008-2022 in both..."

Then set xlim to the desired range in each...it is what controls the range of the x-axis.

In the example code below where I dynamically set the upper range to the first day of the following year of the data in the graph, you could do something like (presuming you had both datasets as intended, not just the same data for both)

xlim([datetime(min(year(T3(1)),year(T5(1))),1,1),datetime(max(year(T3(1)),year(T5(1)))+1,1,1])

which would return the min and max years between the two datasets and make the same adustment to the beginning of first and subsequent years. Or, of course, you can always hardcode in exact limits wanted if there are specific points in time of interest, not just the overall range of the data and to make even/same axes on each plot. You might also want to explore <linkaxes> to put the two in synch with each other if you choose to interactively zoom or somesuch.

"...why I should use datetime instead"

The short answer is "just because!". <VBG> While tongue-in-cheek quip, the features and comparative ease of use of the datetime class as compared to datenum are simply not to be ignored (besides the note in the documentation from TMW that reads <datenum "(Not recommended; use datetime or duration)">. In particular, for the specific use here, the plot() axes are now datetime aware and you can thus get rid of the terrible, klunky now outdated datetick routine entirely and set formats, dateticks, limits, etc.,etc., etc., ... directly with datetime values instead. Yes, you can set xlim with datenum with an analogous call to datenum passing year,month,day, but the supporting routines for the venerable datenum are not as convenient in many cases and even if are for a particular case, you're still stuck with datetick which is truly to be a thing to be glad to be able to do away with.

That doesn't even begin to touch the use in timetable and the like...or the fact that TMW has relegated the venerable datenum time storage mechanism to the dustbin and it will continue to get more and more out-of-date with each and every newer release. It won't ever be completely dropped because it's been around for too long and would cause too much grief in existing code, but shouldn't be used for new code unless there's some truly pressing outside need.

Besides the coding benefits, with datetime you get a display format for free along with it; no more need to convert to "real" time with datestr to see if what you have is what you think you got...the display format at the command line and in tables, etc., etc., is already there in readable date/time units (that you can change to display in your preferred format for the particular use/place at will). You simply cannot do that with a datenum

fn2=websave('compound_tide_1516_data_wl_5.mat','https://www.mathworks.com/matlabcentral/answers/uploaded_files/1134780/compound_tide_1516_data_wl_5.mat');
whos('-file',fn2)
  Name                      Size            Bytes  Class     Attributes

  moving_K1_ht_wl_3         1x5                40  double              
  moving_M2_ht_wl_3         1x5                40  double              
  moving_M4_ht_wl_3         1x5                40  double              
  moving_O1_ht_wl_3         1x5                40  double              
  moving_S2_ht_wl_3         1x5                40  double              
  moving_SA_ht_wl_3         1x5                40  double              
  time_point_1516_wl_3      1x5                40  double              

The second .mat file still contains the same 3-yr data as the first...so can't illustrate w/o taking the time to make up sample data...

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

dpb 2022-9-24

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1811230-how-timelines-unify-in-subplot#answer_1060035

fn1=websave('compound_tide_1516_data_wl_3.mat','https://www.mathworks.com/matlabcentral/answers/uploaded_files/1134780/compound_tide_1516_data_wl_3.mat');

fn1

fn1 = '/users/mss.system.xVSnbh/compound_tide_1516_data_wl_3.mat'

whos('-file',fn1)

Name Size Bytes Class Attributes moving_K1_ht_wl_3 1x5 40 double moving_M2_ht_wl_3 1x5 40 double moving_M4_ht_wl_3 1x5 40 double moving_O1_ht_wl_3 1x5 40 double moving_S2_ht_wl_3 1x5 40 double moving_SA_ht_wl_3 1x5 40 double time_point_1516_wl_3 1x5 40 double

load(fn1)

fn2=websave('compound_tide_1516_data_wl_5.mat','https://www.mathworks.com/matlabcentral/answers/uploaded_files/1134780/compound_tide_1516_data_wl_5.mat');

fn2

fn2 = '/users/mss.system.xVSnbh/compound_tide_1516_data_wl_5.mat'

whos('-file',fn2)

Name Size Bytes Class Attributes moving_K1_ht_wl_3 1x5 40 double moving_M2_ht_wl_3 1x5 40 double moving_M4_ht_wl_3 1x5 40 double moving_O1_ht_wl_3 1x5 40 double moving_S2_ht_wl_3 1x5 40 double moving_SA_ht_wl_3 1x5 40 double time_point_1516_wl_3 1x5 40 double

%return

%load(fn2)

clear fn*

T3=datetime(time_point_1516_wl_3,'ConvertFrom','datenum');

W3=[moving_M2_ht_wl_3;

moving_S2_ht_wl_3;

moving_K1_ht_wl_3;

moving_O1_ht_wl_3].';

% T5=datetime(time_point_1516_wl_5,'ConvertFrom','datenum');

% W5=[moving_M2_ht_wl_5;

% moving_S2_ht_wl_5;

% moving_K1_ht_wl_5;

% moving_O1_ht_wl_5;

% moving_SA_ht_wl_5].';

C={'r';'b';'c';'m';'k'}; % colors; can't use multiple named colors in 'color'

hAx(1)=subplot(2,1,1);

hL(1,:)=plot(T3,W3,'--*');

xlim([datetime(year(T3(1)),1,1),datetime(year(T3(end))+1,1,1)])

set(hL(1,:),{'Color'},C(1:numel(hL(1,:))));

title('WL:3 yr')

hAx(2)=subplot(2,1,2);

hL(2,:)=plot(T5,W5,'linestyle','--*');

Unrecognized function or variable 'T5'.

xlim([datetime(year(T5(1)),1,1),datetime(year(T5(end))+1,1,1)])

set(hL(2,:),{'Color'},C(1:numel(hL(2,:))));

title('WL:5 yr')

Recasts into more compact array form instead of all the terribly-long-and-verbose-variable-names and recasts the time data into datetime. I also then expanded the axes limits to the beginning the year of the first time point and then to the next year after the last year.

NB: The data in the two files is the same 3-YR data so can't do anything about the 5-yr data to see what the underlying Q? really was...if the above doesn't provide sufficient hints on progressing, replace the 5-yr data .mat file with the correct data and explain the problem more precisely.