Plot a vector that contain dates in concatenated number format and plot it using dates

Hello all, I am quite a new user,then sorry if the question is too basic. I got a vector [20000101.1 20000101.2 20000101.3.......20000912.15] the first 4 digits correspond to the year, the second 2 to the month, the third 2 to the day and after the dot is the time in hrs (no minutes, no seconds). Now I would like to plot some values corresponding to these dates where the x axis is displaying the date and time say ( Jan 01 2000, 1 AM...etc). Any idea or advice would be higly appreciated. Thank you very much

Wilson

Answer 1

[Andrei Bobrov]

A = [[20000101.1 20000101.2 20000101.3 20000912.15]];
d = datenum(num2str(A'),'yyyymmdd.HH');
out = datestr(d,'mmm dd yyyy, HH PM');

corrected

A = [20000101.01 20000101.03 20000912.15 20000101.2];
k = strjust(num2str(A'),'left');
k(k == ' ')='0';
out = datestr(datenum(k,'yyyymmdd.HH'),'mmm dd yyyy,HHPM')

[Jan]

Do you have any reasons for the doubled square brackets?

Your method cannot distinguis 20000101.1 and 20000101.10.

[Oleg Komarov]

To ensure that the decimals won't run away on their own :)

[Jan]

I'm curious how Wilson will handle the ambiguos format.

Answer 2

[Fangjun Jiang]

a=[20000101.1 20000101.2 20000101.3 20000912.15];
b=num2str(a');
c=datenum(b,'yyyymmdd.HH');
h=axes;
plot(c,1:4);
d=get(h,'xtick');
e=datestr(d,'mmm dd yyyy HH PM');
set(h,'xticklabel',e)

Answer 3

[Jan]

There is no difference between 20000101.1 and 20000101.10 ?! So if your data are really double values, I assume the best would be to create the data again from scratch using a unique format.

[Oleg Komarov]

Well, it is easy!

20000101.1 --> 2000 01 01 01

20000101.10 --> 2000 01 01 10