find duplicate values using accumarray() function:

I have data that contains longitude,latitude,access point address and signal level. I would like to find duplicate longitude,latitude,access point address, then average the signal level over them, and then replacing the original rows with the new single row. How to do this?

For example (made up data):

Starting data:

if true
33.121,-70.630,aa:91:52:9c:72:21,-80
33.122,-70.631,bb:91:52:9c:72:21,-75
33.123,-70.632,cc:91:52:9c:72:21,-84
33.124,-70.633,dd:91:52:9c:72:21,-57
33.121,-70.630,aa:91:52:9c:72:21,-84
33.122,-70.631,bb:91:52:9c:72:21,-73
end

Would become

if true
33.121,-70.630,aa:91:52:9c:72:21,-82
33.122,-70.631,bb:91:52:9c:72:21,-74
33.123,-70.632,cc:91:52:9c:72:21,-84
33.124,-70.633,dd:91:52:9c:72:21,-57
end

[the cyclist]

How are your data currently stored (e.g. what data type)? Can you post the data, or a small sample that has duplicates?

Answer 1

[Razvan Carbunescu]

Using accumarray specifically you could do this by combining with findgroups :

[idx,unique_long,unique_lat,unique_acc] = findgroups(long,lat,acc);
[unique_signal] = accumarray(signal,idx,[],@mean);

In R2018a could also use groupsummary but the data should be in a table:

T = table(long,lat,acc,signal);
GT = groupsummary(T,{'long','lat','acc'},'mean','signal')

[Nadir Adam]

Thank you very much :)