Finding the cell indices of a cell contains a specific text

Hi, I am working with a large CSV file containing blood pressure data (approx 14 million datapoints). I managed to import the data successfully, creating 2 cell arrays (NUMBERS and TEXT). One contains the actual numbers that I want to use for my calculations (NUMBERS), the other (TEXT) contains the information (as a string) about the associated time-points. Since NUMBERS and TEXT run parallel, I can use the cell indices from TEXT, to get to the data in NUMBERS for further computation.

What I am trying to figure out is how to find a specific string in TEXT; e.g. what are the indices of a time point (e.g. 12:00:00)? I started to work with 'strfind' but this only gives me either 0 or 1; not the cell indices that I need. Moreover, the strings in TEXT contain 'Value, Time, Date' (eg. '103.9685,10:27:37,2016-02-03') which means that I only need the middle portion '10:27:37'of the string.

Can anyone provide me with insights/directions as to what strategy/method is best to use in this case?

Thanks! Barry

Answer 1

[JESUS DAVID ARIZA ROYETH]

I can help you with what you need in this example, it will give you the indices you need :

TEXT={'103.9685,10:27:37,2016-02-03','103.9685,12:00:00,2016-01-03' '103.9685,11:27:37,2016-02-03' '103.9685,12:00:00,2016-02-03' };
indices=find(cell2mat(cellfun(@(x) ~isempty(strfind(x,'12:00:00')),TEXT,'UniformOutput',false)));

Answer 2

[B Janssen]

Hello Jesus, Thanks for your swift answer. It works! Meanwhile after a strong coffee, I came up with a slightly different solution which also seems to do the trick.

Basically it uses the same example array (TEXT), but subsequently uses an extra step (ARRAY) to find the INDICES:

 % 
TEXT={'103.9685,10:27:37,2016-02-03','103.9685,12:00:00,2016-01-03' '103.9685,11:27:37,2016-02-03' '103.9685,12:00:00,2016-02-03' };
 %
ARRAY = cellfun(@(x)~isempty(strfind(x,'12:00:00')), TEXT);
 %
INDICES = find(ARRAY(:) > 0);

The INDICES determine in which row(s) the string(s) can be found. I have timed the code to see which was faster; it turns out that my approach, looking for one string in 14,9915,305 datapoints took 100.15 seconds; while your's took 119.96 seconds. Not exactly sure why this is; but for large data sets a time difference of 20% can be significant.

Again, thanks for your swift and effective response. I hope that combined with my part, others may find this helpful as well!

Barry