Sum daily values for every month in matlab

Question

I have got a matrix with values in.

The first column of the matrix is the date in the following form, 19260701 pr YYYYMMDD.

The other columns of the matrix are series.

19260702    0.026   0.000   NaN 1.175
19260706    0.009   0.000   NaN 1.842
19260707    1.388   0.001   NaN 9.061
19260708    1.147   0.028   NaN 0.067
19260709    0.604   0.018   NaN 0.000
19260710    7.255   0.020   NaN 0.005
19260712    0.085   0.093   NaN 1.832
19260713    0.163   0.025   NaN 3.897
19260714    1.294   0.545   NaN 0.188
19260715    0.256   0.077   NaN 0.001
19260716    0.001   0.002   NaN 0.018
19260717    0.000   0.015   NaN 1.863
19260719    0.002   0.062   NaN 1.465
19260720    2.761   0.028   NaN 6.453
19260721    1.998   0.067   NaN 0.328
19260722    0.160   0.123   NaN 0.651
19260723    0.009   0.000   NaN 0.001
19260724    0.005   0.000   NaN 0.000
19260726    0.016   0.002   NaN 0.860
19260727    0.022   0.000   NaN 0.329
19260728    0.002   0.001   NaN 0.857
19260729    0.000   0.343   NaN 2.125
19260730    0.002   0.001   NaN 1.265
19260731    0.000   0.000   NaN 0.283
19260802    0.000   0.010   NaN 0.815
19260803    0.000   1.020   NaN 27.701
19260804    0.000   0.197   NaN 4.162
19260805    0.027   0.016   NaN 42.120
19260806    0.046   0.200   NaN 15.163
19260807    0.284   0.004   NaN 0.382
19260809    1.330   0.000   NaN 3.102
19260810    1.066   0.016   NaN 0.035
19260811    0.261   0.119   NaN 0.249
19260812    0.014   0.031   NaN 328.139
19260813    0.024   0.042   NaN 40.248
19260814    0.094   0.047   NaN 1.460
19260816    0.042   0.007   NaN 25.928

Is it possible to Sum the values in each column of the matrix based on the month?

Apologise,

Following the comments,

I dont just want to sum by month, but each yearmonth, i.e sum Jan 1960, Feb 1960 etc.

Answer 1

您可以使用unique获取每个月的标签，然后splitapply基于月份标签在其他列中累积值（accumarray可以，但仅在一个列上有效）。

如果考虑同一个月的 2018年4月和2019年4 月：

month = mod(floor(data(:,1)/100-1), 100)+1;
[month_name, ~, month_id] = unique(month);
result = splitapply(@sum, data(:,2:end), month_id);
result_with_month = [month_name result];

如果考虑2018年4月和2019年4月的不同月份：

month = floor(data(:,1)/100);
[month_name, ~, month_id] = unique(month);
result = splitapply(@sum, data(:,2:end), month_id);
result_with_month = [month_name result];

提供的数据的示例结果：

result =
   1.0e+02 *
   0.172050000000000   0.014510000000000   NaN   0.345660000000000
   0.031880000000000   0.017090000000000   NaN   4.895040000000000
result_with_month =
   1.0e+05 *
   1.926070000000000   0.000172050000000   0.000014510000000   NaN   0.000345660000000
   1.926080000000000   0.000031880000000   0.000017090000000   NaN   0.004895040000000

Answer 2

这是一个典型的示例，您可以在其中使用表而不是矩阵。

A = [19260702    0.026   0.000   NaN 1.175
     ...
     19260816    0.042   0.007   NaN 25.928];

% Matrix to table conversion
T = array2table(A,'VariableNames',{'DATE','S1','S2','S3','S4'});
% We add a new column "month"
T.MONTH = floor(T.DATE/100);
% varfun can apply a custom function to your table and group the result according 
% to one (or more) variable(s)
Result = varfun(@sum,T,'InputVariables',{'S1','S2','S3','S4'},'GroupingVariables','MONTH')

懒家伙选项：

如果您很懒惰并且不想手动添加变量的名称，则还可以使用线性索引：

T = array2table(A) %column name will be 'A1','A2',....
T.MONTH = floor(T{:,1}/100);
Result = varfun(@sum,T,'InputVariables',2:5,'GroupingVariables','MONTH')

高尔夫代码选项

如果您喜欢短代码，则可以缩短参数名称。例如，if InputVariables是唯一以开头的参数名称，I您可以简单地使用I代替InputVariables。

Result = varfun(@sum,T,'I',2:5,'G',6)