Iterating through Matlab Table

I have a simple Matlab table with dates and 2 additional columns that contain numbers and strings. The dates are NOT daily, that is sometimes multple days are missing.

I would like to forward fill the table such that the table becomes a daily table and for each day that is missing, the rows from the previous date are copied.

Answer 1

[KSSV]

LEt T be your table. With first column as dates.

% First make the dates daily 
thedates = T.(1) ;  % existing dates of the table 
thedates_new = (thedates(1):thedates(2))' ; %  Make daily dates 
% GEt the existing dates indices 
[idx,ia] = ismember(thedates_new,thedates) ; 
C2 = T.(2) ; % second column of table 
C2_new = NaN(size(thedates_new)) ;
C2_new(idx) = C2;   
C2_new = fillmissing(C2_new) ; 
T_new = table(thedates_new,C2_new) ; 

Alternatively you may have a look on the function retime

[Metin Akyol]

Thank you so much for you help!!

Answer 2

[Seth Furman]

timetable has a retime method that supports resampling by copying the previous value.

Resample or aggregate data in timetable, and resolve duplicate or irregular times - MATLAB retime (mathworks.com)

tt = timetable(datetime(2020,1,[1;3;4;7;9]), [1;3;4;7;9])

tt = 5×1 timetable
       Time        Var1
    ___________    ____

    01-Jan-2020     1  
    03-Jan-2020     3  
    04-Jan-2020     4  
    07-Jan-2020     7  
    09-Jan-2020     9  

retime(tt, "daily", "previous")

ans = 9×1 timetable
       Time        Var1
    ___________    ____

    01-Jan-2020     1  
    02-Jan-2020     1  
    03-Jan-2020     3  
    04-Jan-2020     4  
    05-Jan-2020     4  
    06-Jan-2020     4  
    07-Jan-2020     7  
    08-Jan-2020     7  
    09-Jan-2020     9  

[Metin Akyol]

Thank you so much for you help!!

Answer 3

[Arif Hoq]

date = {'2014-05-20';'2014-05-21';'2014-05-22';'2014-05-24';'2014-05-25'};  % define the date
name = {'Sanchez';'Johnson';'Li';'Diaz';'Brown'};  % define the string 
age = [38;43;38;40;49];  % define the numerical
T=table(date,name,age)  % make a table

T = 5×3 table
         date            name        age
    ______________    ___________    ___

    {'2014-05-20'}    {'Sanchez'}    38 
    {'2014-05-21'}    {'Johnson'}    43 
    {'2014-05-22'}    {'Li'     }    38 
    {'2014-05-24'}    {'Diaz'   }    40 
    {'2014-05-25'}    {'Brown'  }    49 

idx = diff(datenum(T.date, 'yyyy-mm-dd')).'==1;  % return the missing date
idx2=find(idx==0);   % missing date index
newdate=[T.date(1:idx2);T.date(idx2);T.date(idx2+1:end)]; % concatenate the date
newname=[T.name(1:idx2);T.name(idx2);T.name(idx2+1:end)];  % concatenate the name
newage=[T.age(1:idx2);T.age(idx2);T.age(idx2+1:end)]; % concatenate the age
newTable=table(newdate,newname,newage)  % newtable

newTable = 6×3 table
       newdate          newname      newage
    ______________    ___________    ______

    {'2014-05-20'}    {'Sanchez'}      38  
    {'2014-05-21'}    {'Johnson'}      43  
    {'2014-05-22'}    {'Li'     }      38  
    {'2014-05-22'}    {'Li'     }      38  
    {'2014-05-24'}    {'Diaz'   }      40  
    {'2014-05-25'}    {'Brown'  }      49  

[Metin Akyol]

Thank you so much for you help!!