How to omit milliseconds from a table.

Omitting milliseconds when importing datetime.

Hello everyone, I am very new to matlab and writing scripts so please be patient with me.

I am trying to omit the milliseconds from the time I am importing from a CVS file. I came up with the below code to import cvs files based on a naming prefix into a table, but "DatetimeFormate" only changes the displaying of the values not how it is stored.

Ultimately I want to filter out rows from the table with the same HH:mm:ss and join this table with a second one based on the time column. I tried to filter for unique values, or calculating the mean of the values with the same time, but every entry is kept because of the milliseconds being different, so I am looking for a way to omit them.

fSearch=fullfile('TC*'); % build matching wildcard expression
d_TC = dir(fSearch); % and look for match
opts_TC = detectImportOptions(d_TC.name,"VariableNamingRule","preserve"); %Preserving Column Names
opts_TC = setvaropts(opts_TC,"Time","Type","datetime",DatetimeFormat="HH:mm:ss"); %InputFormat="MM/dd/yyyy HH:mm:ss.SSS"
TC = readtable(d_TC.name,opts_TC);

The Resulting table is 90394x6 with the first column being datetime Values like "'11:52:54" and the following five double values for temperatures from a sensor.

Thank you!

Answer 1

[Walter Roberson]

dateshift

TC.Time = dateshift(TC.Time, 'start', 'second');

[Les Beckham]

FYI, you might want to add the 'nearest' rule argument to specify rounding instead of truncation:

TC.Time = datetime(2023, 11, 6, 2, 49, 58.8)

TC = struct with fields:
    Time: 06-Nov-2023 02:49:58

TC.TimeTruncated = dateshift(TC.Time, 'start', 'second')

TC = struct with fields:
             Time: 06-Nov-2023 02:49:58
    TimeTruncated: 06-Nov-2023 02:49:58

TC.TimeRounded = dateshift(TC.Time, 'start', 'second', 'nearest')

TC = struct with fields:
             Time: 06-Nov-2023 02:49:58
    TimeTruncated: 06-Nov-2023 02:49:58
      TimeRounded: 06-Nov-2023 02:49:59

[Constantin]

Thank you! This is exactly what I needed. I did end up going with the rounded time, as that will be more accurate when I merge it with the other table.