Replacing NaN with the average of previous and next cell value

5 次查看（过去 30 天）
krai
krai 2018-6-7
评论： Jan 2018-6-8
I have a data set like,
time 1 2 3 4 5 6 7 8
exp1 45 67 78 NaN 80 81 82 83
These are two separate column with time and exp1 as header. how do I replace the NaN with the average of previous cell value and the next. Like in the above example, NaN should be replaced by 79 ((78+80)/2 )
Thanks in advance
  3 个评论
显示 1更早的评论
Jan
Jan 2018-6-7
Please post the input data in valid Matlab syntax. A "data set" might be a variety of things.

请先登录，再进行评论。

请先登录，再回答此问题。

采纳的回答

Razvan Carbunescu
This sounds like a use for fillmissing with linear option:
>> exp1 = [45 67 78 NaN 80 81 82 83];
>> fillmissing(exp1,'linear')
ans =
45 67 78 79 80 81 82 83

更多回答（2 个）

Jan
Jan 2018-6-7
编辑：Jan 2018-6-7
Maybe:
y = fillgaps(exp1, 3, 1)
% Needs Matlab >= R2016a
[EDITED] Sorry, this does not do, what I expect.
According to the documentation I'd expect a linear fit to the neighboring
elements.
Or:
valid = ~isnan(exp1);
y = exp1;
y(~valid) = interp1(time(valid), exp1(valid), time(~valid))
This is a linear interpolation.
Paolo
Paolo 2018-6-7
编辑：Paolo 2018-6-7
If you are working with a table:
%Table.
var = {'time';'exp1'};
time = [1;2;3;4;5;6;7;8];
exp1 = [45;67;78;NaN;80;81;82;83];
t = table(time,exp1);
x = t.exp1;
nan = isnan(x);
indx = find(nan);
x(nan) = (x(indx-1)+x(indx+1))/2;
t.exp1 = x;
If you are working with a cell array:
%Cell array.
var = {'time',1,2,3,4,5,6,7,8;'exp1',45,67,78,NaN,80,81,82,83};
x = cell2mat(var(2,2:end));
nan = isnan(x);
indx = find(nan);
x(nan) = (x(indx-1)+x(indx+1))/2;
var(2,2:end) = num2cell(x);

类别

在 Help Center 和 File Exchange 中查找有关 NaNs 的更多信息

标签

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by