Calculate mean of numeric column vector based on repeated rows in a string array

Hello, I have a string array (227293-by-3) consisting of weekly dates, latitude and longitude as in the following example:

"20101024 _ 20101031"	  "416"	"826"
"20101024 _ 20101031"	  "416"	"826"
"20101024 _ 20101031"	  "415"	"826"
"20101024 _ 20101031"	  "415"	"827"
"20101024 _ 20101031"	  "415"	"827"
"20101024 _ 20101031"	  "415"	"827"
"20101024 _ 20101031"	  "415"	"827"

I need to calculate the mean of the values in a numerical column vector (227293-by-1) corresponding to the repeating identical rows in the previous string array.

The desired output as per the example would be a column vector:

[mean of rows 1 and 2

row 3

mean of rows 4 through 7...]

I could get the index of the beginning of unique rows in the string array but how could I use that or if there is another method.

Thanks for any assistance.

Answer 1

[dpb]

First, don't use a string array to hold the numeric data -- you don't show us how you got to the above, but if it's coming from a file, use readtable to import it instead of whatever you did...

Given that, then use

tD=array2table(data,"VariableNames",{'Date Range','Lat','Lon'});
tD.("Date Range")=categorical(tD.("Date Range"));
tD.Lat=str2double(tD.Lat);
tD.Lon=str2double(tD.Lon);
tD.SomeVariable=randi(100,size(tD.Lat))
tGrpMean=groupsummary(tD,{'Lat','Lon'},'mean','SomeVariable')
tD =
  7×4 table
        Date Range         Lat    Lon    SomeVariable
    ___________________    ___    ___    ____________
    20101024 _ 20101031    416    826         38     
    20101024 _ 20101031    416    826         60     
    20101024 _ 20101031    415    826          6     
    20101024 _ 20101031    415    827         44     
    20101024 _ 20101031    415    827         16     
    20101024 _ 20101031    415    827         84     
    20101024 _ 20101031    415    827         55     
tGrpMean =
  3×4 table
    Lat    Lon    GroupCount    mean_SomeVariable
    ___    ___    __________    _________________
    415    826        1                   6      
    415    827        4               49.75      
    416    826        2                  49      
>> 

using the data above and working around the unfortunate choice of using string array...