interpolate the duplicate values

Hi

Imagine I have vector of datetime values in increaing (not strictly increasing) order, such as follows:

'14-11-2022 05:18:14.000'

'14-11-2022 05:18:14.000'

'14-11-2022 05:18:14.000'

'14-11-2022 05:18:15.000'

'14-11-2022 05:18:15.000'

'14-11-2022 05:18:16.000'

'14-11-2022 05:18:17.000'

I want to convert every slice of duplicate values into an evenly-spaced slice, so the above vector becomes:

'14-11-2022 05:18:14.000'

'14-11-2022 05:18:14.333'

'14-11-2022 05:18:14.666'

'14-11-2022 05:18:15.000'

'14-11-2022 05:18:15.500'

'14-11-2022 05:18:16.000'

'14-11-2022 05:18:17.000'

I know how to do it with loops, just wondering if there's a clever way to do it, probably using accumarray and histcounts.

thanks

[Rik]

Before you start accusing posts of being generated, why don't you go through the trouble of trying something yourself alrready? You can start by providing your data as actual data.

dates={...
    '14-11-2022 05:18:14.000'
    '14-11-2022 05:18:14.000'
    '14-11-2022 05:18:14.000'
    '14-11-2022 05:18:15.000'
    '14-11-2022 05:18:15.000'
    '14-11-2022 05:18:16.000'
    '14-11-2022 05:18:17.000'};
% Example input vector of datetime values 
dates = datetime(dates,'InputFormat','dd-MM-yyyy HH:mm:SS.000')

dates = 7×1 datetime array
   14-Nov-2022 05:18:00
   14-Nov-2022 05:18:00
   14-Nov-2022 05:18:00
   14-Nov-2022 05:18:00
   14-Nov-2022 05:18:00
   14-Nov-2022 05:18:00
   14-Nov-2022 05:18:00

There is probably a clever way to do this, but it will likely help if you show what you tried.

And can you confirm the constraint that each run of duplicate values will end at the next second? In that case you can probably use some run length encoding tricks.

[mahdi Babayi semiromi]

Hi

Yes each run of duplicate values will end at the next second.

thanks

[Rik]

Well, what did you try? Have you looked into run length encoding?

Answer 1

[Gayatri Rathod]

Hi mahdi,

you can use accumarray and histcounts to achieve this in MATLAB. Here's one way to do it:

1.Convert your date values to datetime format and create input vector of it (e.g., using datetime).

% Example input vector of datetime values 
inputVec = datetime(dates); 

2. Calculate the differences between adjacent serial dates using the diff function:

diffs = diff( inputVec); 

3. Use histcounts to find the indices of the duplicate values:

[~, edges, bin] = histcounts(diffs); 
idx = find(bin > 1); 

4. use accumarray to add the appropriate fractions of a second to each group of duplicate values

% Syntax 
accumarray(ind,data,[],fun) % applies the function fun to each group in data specified by ind. 
                            % Specify fun using the @symbol. for eg. @sum. 

5.Convert the resulting values back to appropriate datetime format if needed.

The resulting vector should be the evenly spaced version of the input dates.

You can read more about the accumarray, histcounts, datetime, diff and find function from the following documentations: accumarray function, histcounts function, datetime function, diff function, find function.

Hope it helps!  

Regards,

Gayatri Rathod