finding the mean based on a specific value in other column

Guys I have following data as an example.The data contain 4 coloumns.want to average 4th coloumn when 1st couloumn is equal to 527.1235 and third coloumn is 927.5

Answer 1

[Asad (Mehrzad) Khoddam]

You can find the rows with the first condition and the other rows for the second condition. The intersection of the two rows, are the row numbers that satisfy both conditions:

 rows = intersect(find(a(:,1)==527.1235), find(a(:,3)==927.5));
 % average of the above rows
 avg = mean(a(rows,4));
 disp(avg)
 
 

or simpler :

 rows = find(a(:,1)==527.1235 & a(:,3)==927.5);
 % average of the above rows
 avg = mean(a(rows,4));
 disp(avg)

[madhan ravi]

Note this answer doesn't include the floating point comparison. Use of find() for the second approach is not necessary ,a well use of logical indexing is more than enough.

[Najam us Saqib Fraz]

Thank You !! But I get NaN while use both the approaches !! The data I showed is a little part of my huge data...While importing it in Matlab and using your suggested solution it gives me NaN.....

[Asad (Mehrzad) Khoddam]

When data is missing in the data file, it shows as NaN. You can remove the lines that are NaN

use this option:

avg = mean(a(rows,4),'omitnan');

Answer 2

[madhan ravi]

ix = (abs(column_1 - 527.1235) <  1e-4) &...
    (abs(column_1 - 927.5) < 1e-1);
M = mean(column_4(ix))

Answer 3

[Jon]

Lets say you have put your data into an array X.

Find a logical index where the rows match your criteria using:

criteria = [527.1235 927.5]
idl = ismember(X(:,[1,3]),criteria,'rows')

then do the averaging on the 4th colulmn for the rows where the criteria matches

xMean = mean(X(idl,4))

[Jon]

Wow, just while I was posting this two other answers appeared. So you have some choices. If I understand your problem correctly though, then this should be a pretty clean way to do it. I guess if you are just matching on two columns then there are many ways to do it. I think this approach scales well though if at some point you have a similar problem but with multiple columns (or whole rows) you want to match on.

[madhan ravi]

>> X

X =

  527.1235    1.0000  927.5000

>> criteria

criteria =

  527.1235  927.5000

>> idl = ismember(X(:,[1,3]),criteria,'rows')

idl =

  logical

   0

>> 

[Najam us Saqib Fraz]

Thank You !! But I get NaN while use both the approaches !! The data I showed is a little part of my huge data...While importing it in Matlab and using your suggested solution it gives me NaN...

[Asad (Mehrzad) Khoddam]

When data is missing in the data file, it shows as NaN. You can remove the lines that are NaN

[Jon]

I think you are trying to show that floating point comparisons could be a problem if they are not exact. I assumed they were exact, in any case given your example I get idl = 1 not 0

>> X = [527.1235    1.0000  927.5000],criteria =[527.1235  927.5000]

X =

  527.1235    1.0000  927.5000


criteria =

  527.1235  927.5000

>> idl = ismember(X(:,[1,3]),criteria,'rows')

idl =

  logical

   1