Input data file containing dashes, colon and slashes

I have a data file which contains data in the following format :

 2020-Mar-08 02:28     09 46 28.67 +17 21 52.5 -12.22   3.95 0.00242027652616  -0.0482121 155.1431 /T  24.7981
 2020-Mar-09 00:00     10 45 35.20 +12 42 10.2 +12.58   3.65 0.00239787063143  -0.0287020 168.5361 /T  11.4363
 2020-Mar-10 23:34     11 43 12.33 -07 10 19.6 -12.73   3.52 0.00238767904556  -0.0063201 173.8507 /L   6.1347
 2020-Mar-11 19:21     12 39 32.82 +01 11 06.2 -12.40   3.82 0.00239061880814   0.0163142 161.3156 /L  18.6406
 2020-Mar-12 12:07     13 35 08.81 -04 49 28.4 +12.03   4.12 0.00240606204543   0.0365759 147.4799 /L  32.4459

There is a space at the very first column (before all 2020s). There are five spaces between the time and the next quantity and three spaces between the 9th and 10th quantity. The sixth quantity always has a + or a - sign, but the 12th quantity does not contain a + sign when it is above 0.

How to read this file for the dates to go in one vector, the times in another vector (without the colon) and the rest of the quantities in separate vectors ? The /T needs to be replaced with digit '1' and /L with digit '2'.

[Rik]

What have you tried? It is easier to improve slow working code than to come up with new code.

Answer 1

[Walter Roberson]

filename = 'data.txt';
opt = detectImportOptions(filename);
opt.VariableTypes{2} = 'duration';
opt.VariableOptions(2).InputFormat='hh:mm';

t = readtable(filename, opt);

dates = t.Var1;
times = t.Var2;
[and keep going with the other .Var up to 15]
[But really it is better to leave everything in the table]

[~, TL] = ismember(t.Var14, {'/T', '/L'});   %TL will be 1 or 2, or 0 if not found

[b]

Everything is ok. Time is ok but it still has this format :

times = 
  5×1 duration array
   02:28
   00:00
   23:34
   19:21
   12:07

It will be nice for it to be in the decimal format where 02:28=02.47 and 19:21=19.35.

[Stephen23]

Try https://www.mathworks.com/help/matlab/ref/duration.hours.html

Unfortunately the datetime Format does not seem to support decimal fractions of an arbitrary unit. With that in mind, you can add the duration times to the datetimes to get complete datetime timestamps.

[Walter Roberson]

times = hours(t.Var2)

However, in my experience it is much more common to want to skip that and to instead go directly to

datetimes = t.Var1 + t.Var2;
datetimes.Format = 'uuuu-MMM-dd hh:mm';

[b]

Many thanks ! Completely works !

Answer 2

[b]

Huh, What happened ??!!

Where did Stephen Cobeldick's code snippet vanish ??!

Stephen - after putting the time specifier as %{hh:mm}T, the code snippet that you suggested gave a very strange result. The time comes correct, but all the rows where the 14th variable is /T are ignored and the table only stores the rows where the value of this variable is /L. So in this case, the table only reads in the bottom three rows, and ignores the top two rows.

Incidentally, how can we remove the colon from the time data so that it can be stored in the decimal format ?