Read one column from a csv file into MATLAB for calculating

I'd like to read one column from a csv file '2004.csv' which already in the MATLAB path. The column contains 'NaN' value and I want to treat it as empty. The column looks like following:

ArrDelay

NaN

-14

2

14

-6

NaN

11

57

...

-6

NaN

11

57

The variable name is 'ArrDelay' which is the title of the column. The column contains 7009729 records and I'd like to calculate the mean of those records.

How can I make it?

Answer 1

[Andrei Bobrov]

out = nanmean(ArrDelay);

or in MATLAB R2015a

out = mean(ArrDelay,'omitnan'));

[Xiaoqing Zhuang]

How to read that column from '2004.csv' file into matlab workspace without using import? My version is MATLAB R2014a. Thank you so much! :)

[Andrei Bobrov]

ArrDelay = csvread('2004.csv')

[Xiaoqing Zhuang]

Can I use the variable name of 'ArrDelay' to specify the column I want to read? Or I need to use the index of the column (14th)?

I tried ArrDelay = csvread('2004.csv', 1, 14)

I got several error messages:

Error using dlmread (line 138)

Mismatch between file and format string.

Trouble reading number from file (row 1u, field 17u) ==> ORD,CLT,599,7,11,0,,0,0,0,0,0,0\n

Error in csvread (line 47) m=dlmread(filename, ',', r, c);

Error in bigmean (line 2) ArrDelay = csvread('2004.csv', 1, 14);

How can I deal with it?

[Andrei Bobrov]

Please attach your csv-file

[Xiaoqing Zhuang]

The data set is very large and it doesn't allow me to attach here. I can describe it.

The column I need combines string 'NaN' and integers. There are A to AC columns (29 columns). The column needed is 15th. There are 7009729 rows(the first row is title for each column, it is string. Tile for the 15th is 'ArrDelay').

Purpose: calculate mean of that column.

I just tried following codes, it failed.

tic

ArrDelay = csvread('2008.csv', 1, 14, [1,14,7009729,15]);

meanresult = nanmean(ArrDelay);

sprintf('%g.\n',toc)

[Xiaoqing Zhuang]

I attached the sample of the csv file. Please have a look. Thanks in advance!