Countering error while using interp1 function

Hi,

I am using the code below but getting the error and can not make sense of that error can anyone help please: Attached are the files I am working with.Basically, I am trying to use the time stamps available in both files attached and find the common timestamps for both velocity and head data hence plot them for visualization but I think the size of the both files are different that's why error occuring?

Code:

close all; clear all; clc;

% read data into tables

Udat = readtable('F:\3-PIV_Experimental_Data\Outlet_110\Data_LaserSheet_D\Data_PIV\5-AverageFilter_VelocitiesExtration\Point_Velocities\Uy\TimeStamps_Uy.xlsx');

Hdat = readtable('F:\3-PIV_Experimental_Data\Outlet_110\Data_LaserSheet_D\Data_Head\Data_Raw\Head_Timestamps.xlsx');

% assign variables

tu = datetime(Udat.TimeStamp); % convert to datetime array

u = Udat.Uy

th = datetime(Hdat.TimeStamp);% convert to datetime array

h = Hdat.Head_mm_; % Matlab converts original heading Head(mm) to this legal variable name

% interpolate values for head at times to match those where we have

% velocity data

hq = interp1(th,h,tu);

% plot the results

plot(u,hq,'-o')

Error:

Error using datetime (line 597)

Numeric input data must be a matrix with three or six columns, or else three, six, or seven separate numeric arrays. You can also create datetimes from a

single numeric array using the 'ConvertFrom' parameter.

Error in headvelocitygraph (line 10)

th = datetime(Hdat.TimeStamp);% convert to datetime array

Answer 1

[John D'Errico]

Your problem is NOT in interp1. It is in the incorrect use of datetime. Read the error message:

Error using datetime (line 597)
Numeric input data must be a matrix with three or six columns, or else three, six, or seven separate numeric arrays. You can also create datetimes from a
single numeric array using the 'ConvertFrom' parameter.

[muhammad choudhry]

yes, thanks for letting me know. How would I correct that error.

[muhammad choudhry]

see why I am not using the same error using the TimeStamps_Uy.xlsx, this file running fine but as soon as code come the file Head_Timestamps.xlsx. it start generating error, both files have the extracted data from the sensors, is it means both are transfering the data in different format?

I read the error it says first change the readtable to read matrix ones I done that it gave me error below,

Dot indexing is not supported for variables of this type.

Error in headvelocitygraph (line 10)

th = datetime(Hdat.TimeStamp,'InputFormat','dd.MM.yyyy HH:mm:ss.SSS');% convert to datetime array

then I used one more line to convert the below line into datetime format but still getting error

th = datetime(Hdat.TimeStamp,'InputFormat','dd.MM.yyyy HH:mm:ss.SSS');% convert to datetime array

[Peter Perkins]

These two files have what appears to be times of day, or more generally, elapsed times. But the two files are apparently formatted completely differently, because readtable reads the timestamps in one as numeric, in the other as text. Neither is ideal; a properly formatted spreadsheet should cause readtable to read them in as timestamps. Anyway:

>> t1 = readtable("Head_Timestamps.xlsx");
Warning: Column headers from the file were modified to make them valid MATLAB identifiers before creating variable names for the table. The original column headers are saved in the VariableDescriptions property.
Set 'VariableNamingRule' to 'preserve' to use the original column headers as table variable names. 
>> head(t1)
    TimeStamp    Head_mm_
    _________    ________
     0.56369         0   
      0.5637       1.2   
     0.56372       1.2   
     0.56373       0.7   
     0.56374       0.8   
[snip]
>> t2 = readtable("TimeStamps_Uy.xlsx");
>> head(t2)
     TimeStamp         Uy    
    ____________    _________
    {'13:32:04'}    0.0023132
    {'13:33:09'}     0.068063
    {'13:33:42'}     0.056888
    {'13:34:14'}     0.070543
    {'13:34:46'}      0.13177
[snip]
>> t1.TimeStamp = days(t1.TimeStamp); t1.TimeStamp.Format = "hh:mm:ss.SSS";
     TimeStamp      Head_mm_
    ____________    ________
    13:31:43.000        0   
    13:31:44.000      1.2   
    13:31:45.000      1.2   
    13:31:45.999      0.7   
    13:31:47.000      0.8   
[snip]
>> t2.TimeStamp = duration(t2.TimeStamp); t2.TimeStamp.Format = "hh:mm:ss.SSS"
t2 =
  45×2 table
     TimeStamp          Uy     
    ____________    ___________
    13:32:04.000      0.0023132
    13:33:09.000       0.068063
    13:33:42.000       0.056888
    13:34:14.000       0.070543
    13:34:46.000        0.13177
[snip]

Answer 2

[Walter Roberson]

th = datetime(Hdat.TimeStamp);% convert to datetime array

You did not use any options when you readtable(): how can you be sure that TimeStamp was read in as either string array or cell array of character vectors? How do you know that the format was not recognized and it might already have been converted to datetime? Or perhaps the column got misidentified as numeric?

[muhammad choudhry]

I saw that, it is reading Hdat not as date but by numbers in matlab as shown below, I am not sure why.

0.563692129629630

0.563703703703704

0.563715277777778

0.563726851851852

0.563738425925926

0.563750000000000

0.563761574074074

0.563773148148148

0.563784722222222

0.563796296296296

0.563807870370370

0.563819444444444

0.563831018518519

0.563842592592593

0.563854166666667

0.563865740740741

0.563877314814815

0.563888888888889

0.563900462962963

0.563912037037037

[Walter Roberson]

The field might be a numeric field that is storing data in terms of full days. 13 hours is about 0.56 days. The mini-viewer on my phone says that there is no date in the field.

[muhammad choudhry]

see did the conversion using the following

th = datetime(Hdat.TimeStamp,'InputFormat','HH:mm:ss.SSS','ConvertFrom','datenum');

code runs file but not plotting anything supposed to plot plot(u,hq,'-o') but I am getting empty graph and whenever I check hq value I am getting

NaN

NaN

NaN

is that mean nothing match ? supposed to be the values which is common in both files

[Walter Roberson]

https://www.mathworks.com/matlabcentral/answers/1670499-converting-the-time-data-into-datetime-format#answer_916444

[Peter Perkins]

I don't think these are datetimes, I think they are durations.

[muhammad choudhry]

You are right Peter, it was durations, but i finally figured that out yesterday by using the class() and updated the code accordingly as shown below. One question since your reply here! how would I know by just looking at the data rather its datetime, or duration ? if the data come out as 12 March 2012 13:03:01 or just 13:03:01. I spent all day yesterday just to understand this

Updated Code:

% read data into tables

Udat = readtable('F:\3-PIV_Experimental_Data\Outlet_110\Data_LaserSheet_D\Data_Head\Data_Raw\Head_TimestampsCSV.csv');

Hdat = readtable('F:\3-PIV_Experimental_Data\Outlet_110\Data_LaserSheet_D\Data_Head\Data_Raw\Head_TimestampsCSV.csv');

%Check the data type before proceeding with the different data =>

u = Udat.Uy

h = Hdat.Head_mm_; % Matlab converts original heading Head(mm) to this legal variable name

tu=duration(Udat.TimeStamp_Uy);

th = duration(Hdat.TimeStamp);

% interpolate values for head at times to match those where we have

% velocity data

hq = interp1(th,h,tu);

% plot the results

plot(u,hq,'-o')

size(h)

size(Hdat)

size(hq)

size(th)

[Peter Perkins]

I can't answer that. You need to have at least some idea what is in your data files. datetime assumes that if you give it a partial (time-only) timestamp, you mean "today". In my opinion, your data files are the culprit and are not formatted properly.