Creating ID using year and event number
古いコメントを表示
Hi,
I have a set of cyclone data that I am trying to preprocess before I can do some stats on it. In order to do this i need to create a unique id for each system using the year it happened and which event in the year it was. In the sample data i have attached, every time Var1 returns to 1, a new system is registered. So Var1(1:3) should have id 1990S1, Var1(4:5) should be 1990S2, Var(6:7) should be 1990S3 and so forth. I want to create a column of Id and put it in Col1 of my time table so that it easy for me to use 'varfun' to do the stats.
Thanks in advance
採用された回答
Try this —
load('sample.mat')
whos('-file','sample.mat')
sample
count = cumsum([1; diff(sample.Var1) < 0]);
ID = year(sample.time) + "S" + count;
sample = addvars(sample, ID, 'Before',1)
The logic is straightforward. The ‘count’ assignment cumulatively adds every instance where the difference in ‘Var1’ decrements, so the result is true or 1. Logic values become numeric when used in calculations, so ‘count’ increments accordingly. The ‘ID’ variable is easy to create using the year and the value for ‘count’ and string variable syntax. (If you do not have string variables, use sprintf to create a similar array.)
I kept all the other variables here. Use removevars to eliminate ‘Var1’ if you no longer want it in your timetablle.
.
4 件のコメント
Sarvesh
2024 年 5 月 27 日
As always, my pleasure!
EDIT — (27 May 2024 at 12:54)
My code runs and produces the correct result here because there is only one year in ‘sample’.
To make it robust to timetable arrays with more than one year, use this variation —
% Create Test Timetable —
time = datetime(reshape(repmat([1990; 1991; 1992], 1, 7).', [], 1), repmat((1:7).',3,1), ones(21,1), 'Format','yyyy-MM-dd');
Var1 = repmat([1;2;1;2;3;1;2],3,1);
T2 = timetable(time,Var1)
[Uyear,idx1,idx2] = unique(year(T2.time)); % Required
Yr_Var1_count = @(Tbl,idx2) accumarray(idx2, (1:numel(idx2)).', [], @(x){cumsum([1; diff(Tbl.Var1(x)) < 0])}); % Separate Vectors For Each Year (Function)
ID = year(T2.time) + "S" + cat(1,cell2mat(Yr_Var1_count(T2,idx2)));
T2 = addvars(T2, ID, 'Before',1)
% Use Ofiginal Timetable —
load('sample.mat')
[Uyear,idx1,idx2] = unique(year(sample.time)); % Required
ID = year(sample.time) + "S" + cat(1,cell2mat(Yr_Var1_count(sample,idx2))); % Separate Vectors For Each Year (Call 'Yr_Var1_count' Function)
sample = addvars(sample, ID, 'Before',1)
NOTE — Each timetable needs to have a separate unique call, as done here. With that and the appropriate arguments to ‘Yr_Var1_count’ it should work with all timetable arrays with the appropriate format.
.
Sarvesh
2024 年 5 月 28 日
As always, my pleasure!
The accumarray approach is efficient as well as fast.
A complete function using this would be:
function newTable = createID(oldTable) % Complete Function
newTable = oldTable;
[Uyear,idx1,idx2] = unique(year(newTable.time)); % Required
Yr_Var1_count = accumarray(idx2, (1:numel(idx2)).', [], @(x){cumsum([1; diff(newTable.Var1(x)) < 0])}); % Separate Vectors For Each Year (Function)
ID = year(newTable.time) + "S" + cat(1,cell2mat(Yr_Var1_count));
newTable = addvars(newTable, ID, 'Before',1);
end
You can either copy it and paste it in a specific script, or save it as createID.m in your MATLAB search path, and then call it when you need it, or both, depending on what you want to do. It does not change the original timetable, creating a new version as the output.
My apologies for not considering the multi-year problem earlier.
Illustrating how to use ‘createID’ —
% Create Test Timetable —
time = datetime(reshape(repmat([1990; 1991; 1992], 1, 7).', [], 1), repmat((1:7).',3,1), ones(21,1), 'Format','yyyy-MM-dd');
Var1 = repmat([1;2;1;2;3;1;2],3,1);
T2 = timetable(time,Var1)
tic
T2new = createID(T2)
toc
% Use Ofiginal Timetable —
load('sample.mat')
sample
tic
sampleNew = createID(sample)
toc
function newTable = createID(oldTable) % Complete Function
newTable = oldTable;
[Uyear,idx1,idx2] = unique(year(newTable.time)); % Required
Yr_Var1_count = accumarray(idx2, (1:numel(idx2)).', [], @(x){cumsum([1; diff(newTable.Var1(x)) < 0])}); % Separate Vectors For Each Year (Function)
ID = year(newTable.time) + "S" + cat(1,cell2mat(Yr_Var1_count));
newTable = addvars(newTable, ID, 'Before',1);
end
.
その他の回答 (0 件)
カテゴリ
ヘルプ センター および File Exchange で Exponents and Logarithms についてさらに検索
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!
Translated by 
