Using RegEx (strrep) on CharArray with For Loop

3 vues (au cours des 30 derniers jours)
Brian Castro
Brian Castro le 29 Avr 2013
Hello. I have a series of dates that need to be reformatted as such:
From:
2007-04-22 01:04:21.000
2007-05-10 01:08:48.000
2007-06-15 22:03:31.000
To:
2007/04/22 01:04:21
2007/05/10 01:08:48
2007/06/15 22:03:31
I currently have this loop to try and replace each "-" with "/". (I also need to cut off the ".000", but I have not gotten to that yet.)
newTimes = get(c, 'START_STR');
for j = 1 : size(newTimes, 1);
fprintf('Looped'); %Just for testing
newsTimes(j) = strrep(newTimes(j), '-' ,'/')
end
However this does not work.
I am no longer getting an error that "Error using ==> strrep Input strings must have one row". Which I had before I put it in a loop. But my output is still wrong.
It does NOT make the swap of '-' for "/" and it also prints out the array for each time the loop runs, as opposed to printing one line for each run of the loop.
Any ideas? Ideally I would like to find a function similar to strrep that can be used in an array... otherwise a way to fix the loop would be great.
My Output looks Like this:
Looped
newstationtimes =
2007-04-22 01:04:21.000
2007-05-10 01:08:48.000
2007-06-15 22:03:31.000
... rest of list
Repeated like 40 times.
  5 commentaires
Daniel Shub
Daniel Shub le 29 Avr 2013
@Cedric just because you don't need regexp, doesn't mean you cannot use it. I bet you would get a vote for a regexp answer.
Cedric
Cedric le 29 Avr 2013
Modifié(e) : Cedric le 29 Avr 2013
Well, I'd personally go for a reshape of the char. stream into a ?x25 array (or 23/24 if \r and or \n are not present), then a replacement like buffer(buffer=='-') = '/', and finally I'd kill columns 20 to 23.
Regexp are useless because of the rigid structure, but if the time format was more flexible or if there was text with variable length between the date and the time, one way to use pattern matching would be:
content = fileread('myFileWithTimeStamps.txt') ;
content_new = regexprep(content, '(.{4})-(..)-(.*?)\.000', '$1/$2/$3') ;

Connectez-vous pour commenter.

Réponse acceptée

Daniel Shub
Daniel Shub le 29 Avr 2013
What about
datestr(['2007-04-22 01:04:21.000';'2007-05-10 01:08:48.000';'2007-06-15 22:03:31.000'], 'yyyy/mm/dd HH:MM:SS')
I think some of the date functions are slow (if so Jan will probably chime in with the faster way) ...
  1 commentaire
Brian Castro
Brian Castro le 29 Avr 2013
Thanks Strongbad!
errr... I mean Daniel.
That worked great. I don't anticipate it being too slow.

Connectez-vous pour commenter.

Plus de réponses (1)

Sean de Wolski
Sean de Wolski le 29 Avr 2013
Modifié(e) : Sean de Wolski le 29 Avr 2013
Two easy ways:
If the data is indeed all column-aligned, how about:
data(:,[5 8]) = '/'
data = data(:,1:end-4);
If the data is not perfectly column aligned:
  • Converting them to date numbers with datenum() and the format that you have?
  • Convert them back to a string using datestr() with your choice output format?
This surely won't be the fastest way.
  1 commentaire
Brian Castro
Brian Castro le 29 Avr 2013
Thanks! Since Im using dates I went with datestr, but both of those should work

Connectez-vous pour commenter.

Catégories

En savoir plus sur Programming dans Help Center et File Exchange

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by