Extract rows from a cell array based on keywords

Hello,

I have created the following cell array:

'6440.016'	'hold'
'7378.016'	'max-pressure-evac 60.0'
'7439.016'	'evac-end'
'8372.016'	'max-pressure-evac 60.0'
'8433.016'	'evac-end'
'9365.016'	'max-pressure-evac 60.0'
'9426.016'	'evac-end'
'10357.016'	'max-pressure-evac 60.0'
'10418.016'	'evac-end'
'11338.016'	'max-pressure-evac 60.0'
'11399.016'	'evac-end'
'12298.016'	'max-pressure-evac 60.0'
'12359.016'	'evac-end'
'13259.016'	'max-pressure-evac 60.0'
'13320.016'	'evac-end'
'14221.016'	'max-pressure-evac 60.0'
'14282.016'	'evac-end'
'15184.016'	'max-pressure-evac 60.0'
'15245.016'	'evac-end'
'16149.016'	'max-pressure-evac 60.0'
'16210.016'	'evac-end'
'17115.016'	'max-pressure-evac 60.0'
'17176.016'	'evac-end'
'18082.016'	'max-pressure-evac 60.0'
'18143.016'	'evac-end'
'19050.016'	'max-pressure-evac 60.0'
'19111.016'	'evac-end'
'19424.016'	'hold-end'
'19553.016'	'hold'
'20272.016'	'max-pressure-evac 60.0'
'20333.016'	'evac-end'
'21109.016'	'max-pressure-evac 60.0'
'21170.016'	'evac-end'
'21919.016'	'hold-end'
'22048.016'	'hold'
'22651.016'	'max-pressure-evac 60.0'
'22712.016'	'evac-end'
'23374.016'	'max-pressure-evac 60.0'
'23435.016'	'evac-end'
'23489.016'	'hold-end'
'23619.016'	'hold'
'24131.016'	'max-pressure-evac 60.0'
'24192.016'	'evac-end'
'24762.016'	'max-pressure-evac 60.0'
'24823.016'	'evac-end'
'25060.016'	'hold-end'
'25190.016'	'hold'
'25640.016'	'max-pressure-evac 60.0'
'25701.016'	'evac-end'
'26212.016'	'max-pressure-evac 60.0'
'26273.016'	'evac-end'
'26631.016'	'hold-end'

Clearly it can be seen that there are multiple occaisons (5 to be exact) where the lines 'max-pressure-evac 60.0' and 'evac-end' alternate with eachother more than once. When this happens I would ideally like to extract the numbers corresponding to the second to last 'evac-end'/'max-pressure-evac 60.0' combo. For example, in:

'25190.016'	'hold'
'25640.016'	'max-pressure-evac 60.0'
'25701.016'	'evac-end'
'26212.016'	'max-pressure-evac 60.0'
'26273.016'	'evac-end'
'26631.016'	'hold-end'

I would want to keep the numbers 25701.016 and 26212.016 which corresond to 'evac-end' and 'max-pressure-evac 60.0' respectively. Any help in solving this problem would be greatly appreciated.

Answer 1

[Jon]

Let say you call your cell array A, then you could do this:

% make first column into numerical values rather than strings
x = str2double(A(:,1));

% make another vector with the event names, just to make easier to work
% with
event = string(A(:,2));

% find instances of first keyword
idl1 = strcmp(event,['max-pressure-evac 60.0']);

% find instance of second keyword
idl2 = strcmp(event,'evac-end');

% mark all of the locations where first keyword is directly followed by
% second keyword
idl = idl1 & [idl2(2:end);0];
idx = find(idl); % get the corresponding indices
% also mark the locations of the second keyword in the group
idl(idx+1) = true;

% find the  end of each group
idx = find(diff(idl) == -1);

% save the second to the last values
evacEndVals = x(idx-2)
maxPressVals = x(idx-1)

with your example data this gives, the following, which I think is what you are asking for.

evacEndVals =

1.0e+04 *

1.8143

2.0333

2.2712

2.4192

2.5701

maxPressVals =

1.0e+04 *

1.9050

2.1109

2.3374

2.4762

2.6212

[Justin Rosenthal]

This works exactly how I needed it to! Thank you so much.

[Jon]

That's great. Good luck with your project.

[Justin Rosenthal]

Thanks again Jon. I know it may be a lot to ask from you but would you know how to handle the following situation:

'111772.016'	'hold'
'127845.016'	'max-pressure-evac 60.0'
'127906.016'	'evac-end'
'133373.016'	'hold-end'
'133502.016'	'hold'
'145343.016'	'max-pressure-evac 60.0'
'145404.016'	'evac-end'
'155103.016'	'hold-end'
'155232.016'	'hold'
'165446.016'	'max-pressure-evac 60.0'
'165507.016'	'evac-end'
'175563.016'	'max-pressure-evac 60.0'
'175624.016'	'evac-end'
'176833.016'	'hold-end'
'176963.016'	'hold'
'186049.016'	'max-pressure-evac 60.0'
'186110.016'	'evac-end'
'195030.016'	'max-pressure-evac 60.0'
'195091.016'	'evac-end'
'198564.016'	'hold-end'
'198695.016'	'hold'
'206837.016'	'max-pressure-evac 60.0'
'206898.016'	'evac-end'
'214911.016'	'max-pressure-evac 60.0'
'214972.016'	'evac-end'
'220296.016'	'hold-end'

where if the 'max-pressure-evac 60.0' does not repeat the value is retained, but if it repeats (as in the earlier example) you again pull out the last 'evac-end'/'max-pressure-evac 60.0' combination?

[Jon]

and if it doesn't repeat do you also pull out the the evac-end that immediately follows the max_pressure-evac 60?

[Jon]

I think this does what you ask:

% make first column into numerical values rather than strings
x = str2double(A(:,1));

% make another vector with the event names, just to make easier to work
% with
event = string(A(:,2));

% find instances of first keyword
idl1 = strcmp(event,['max-pressure-evac 60.0']);

% find instance of second keyword
idl2 = strcmp(event,'evac-end');

% mark all of the locations where first keyword is directly followed by
% second keyword
idl = idl1 & [idl2(2:end);0];
idx = find(idl); % get the corresponding indices
% also mark the locations of the second keyword in the group
idl(idx+1) = true;

% find the  start end of each group
idxStart = find(diff([0;idl]) == 1);
idxEnd = find(diff(idl) == -1);

% mark the number of repeats at the end of each group
rep = zeros(size(idl));
rep(idxEnd) = (idxEnd - idxStart + 1)/2; % each two elements in a group is one repeat

% store the x values that are to be kept in a location corresponding to the
% end of each group
% initialize with nan so we can tell later which have be reassigned
% 
evacEndVals = nan(size(x)); 
maxPressVals = nan(size(x));


% save the values corresponding to single occurences
idx = find(rep==1); % index of end of each single
evacEndVals(idx) = x(idx);
maxPressVals(idx) = x(idx-1);


% save the second to the last values for multiple occurences
idx = find(rep > 1)
evacEndVals(idx) = x(idx-2);
maxPressVals(idx) = x(idx-1);

% get rid of the unassigned values (nan's)
evacEndVals = evacEndVals(~isnan(evacEndVals))
maxPressVals = maxPressVals(~isnan(maxPressVals))

For your new input data set this gives:

evacEndVals =

1.0e+05 *

1.2791

1.4540

1.6551

1.8611

2.0690

maxPressVals =

1.0e+05 *

1.2785

1.4534

1.7556

1.9503

2.1491

If you don't want to retain the evacEndVals when there aren't repeats you can modify these lines

% save the values corresponding to single occurences
idx = find(rep==1); % index of end of each single
evacEndVals(idx) = x(idx);
maxPressVals(idx) = x(idx-1);

to

% save the values corresponding to single occurences
idx = find(rep==1); % index of end of each single
maxPressVals(idx) = x(idx-1);

[Justin Rosenthal]

Thank you again so much Jon. This is again exactly what I was looking to accomplish.