MATLAB: How to manipulate data from a csv datafile containing date, time and number data

time series data analysis

This sounds intuitively easy but it has me puzzled. Columns 1 and 2 contain datetime formatted data, with column 3 double precision numerical. I can import the data using READTABLE with no apparent problem but subsequent manipulation continues to fall over. Is it best to split this into three separate (n x 1) arrays, or can I work with the data as one (n x 3) matrix? My routine is to read in the data, extract portions of the data (e.g. 600 rows) and then statistically analyse the values of the third column. The routine was working fine until I exceeded 24 hours of data, when it then ran into ambiguity due to ambiguous data type. I continue to run into a problem during subsequent code using DATEVEC, which produces an error message regarding data type, and using TABLE2ARRAY which also produces a data-type error. I am using R2016b. I am pretty new to Matlab and have attached a small sample input data. Cheers Bob

Best Answer

f = fopen('Input Data.csv');
c = textscan(f,'%s %s %f','delimiter',',');
fclose(f);
TT = timetable(datetime(strcat(c{1},c{2}),'In','dd-MMM-yyhh:mm:ss a'),c{3},'Va',{'datas'});

Related Solutions

MATLAB: Comparing datetime value from two tables

A datetime always has an hour, minute, second component. The display format of the datetime array may be such that it is not shown, but it is always there. If you've read that array from a text file which didn't have a time component, then it defaults to 00:00:00.

You have several options to compare datetime arrays using just the date

You could set the time component in a to 00:00:00 and use standard comparison functions (e.g. ismember, or join)
You could extract the year, month and day from each with ymd and do array comparison (e.g. with ismember with the 'rows' option)
You could extract the date as strings (using char or datestr) and compare the strings.

Going with option 2, you could do:

ymda = zeros(height(a), 3);
[ymda(:, 1), ymda(:, 2), ymda(:, 3)] = ymd(a.Datetime);
ymdb = zeros(height(b), 3);
[ymdb(:, 1), ymdb(:, 2), ymdb(:, 3)] = ymd(b.Date);
queryresult = b(ismember(ymdb, ymda, 'rows'), :);  %equivalent to select * from b where b.date = a.date

MATLAB: How to import csv-file with headlines (type text) to matrix? – only numbers are needed

Both readtable and readmatrix will handle this file without issue. I would have thought that importdata would delegate to readtable so I'm surprise it didn't work for you.

readtable will label the columns according to the header.

Best Answer

Related Solutions

MATLAB: Comparing datetime value from two tables

MATLAB: How to import csv-file with headlines (type text) to matrix? – only numbers are needed

Related Question