Reorder table variables / columns

402 ビュー (過去 30 日間)
Alan
Alan 2014 年 9 月 18 日
コメント済み: cui,xingxing 2023 年 11 月 17 日
I've created a table from a large csv file using the readtable function. I want to reorder the variables / columns in this table.
For example.
My table's name is CUTable
It has a variable CUTable.Date which is the date the data was taken It is currently column 20 in the table. I want to make it column 2 and shift existing columns 2-19 to 3-20.
It seems there should be a very easy command to do this which I'm missing, but I can't seem to find it anywhere in the help or these forums. I could, but don't want to "brute force" it (e.g. create a new table, and one-by-one copy the variables from the exiting table to the new table in the order I want, then delete the old table) if avoidable.
How can I accomplish this?

採用された回答

Geoff Hayes
Geoff Hayes 2014 年 9 月 18 日
編集済み: Geoff Hayes 2014 年 9 月 18 日
Alan - you can try the following to move the 20th column to the second position
CUTable = [CUTable(:,1) CUTable(:,20) CUTable(:,2:19)];
It isn't as elegant as a simple "move column command" but it should do what you want. In the following example, I inserted the last column in before the second (note that I use end here rather than the last column number)
load patients;
T = table(Age,Diastolic,Gender,Height,LastName,Location,...
SelfAssessedHealthStatus,Smoker,Systolic,Weight)
T = [T(:,1) T(:,end) T(:,2:end-1)]
  3 件のコメント
Ole
Ole 2014 年 10 月 30 日
編集済み: Ole 2014 年 10 月 30 日
You can also use the Workspace browser to modify your table by hand. The useful thing is that Matlab shows you the commands it performs for each action. For example when rearranging columns it does:
T=table; T.x=1; T.y=2; % setup demo table
T=T(:,[2 1]); % rearrange x and y columns
For the above example by Geoff Hayes that would be
T = T(:,[1 end 2:end-1])
which is apparently the way you are supposed to do it.
Peter Perkins
Peter Perkins 2018 年 2 月 21 日
編集済み: Peter Perkins 2018 年 2 月 21 日
Yes, for a simple rearrangement, what Ole suggests is the best way. In some cases, it will be more readable use variable names, something such as
T = array2table(rand(10,3),'VariableNames',{'x' 'y' 'z'})
T = T(:,{'z' 'y' 'x'})
but with more than a handful of variables, probably using numbers is more readable.

サインインしてコメントする。

その他の回答 (2 件)

Eric Sofen
Eric Sofen 2018 年 3 月 28 日
In R2018a, check out the new movevars function for rearranging variables in tables. It's part of a set of new functions for manipulating table variables:

Ben Oeveren
Ben Oeveren 2018 年 2 月 21 日
編集済み: Ben Oeveren 2018 年 2 月 21 日
Or you can use:
oldvariables = T.Properties.VariableNames;
newvariables = {'date','Title','Website'};
[~,LOCB] = ismember(oldvariables,newvariables);
newTable = T(:,LOCB);
  1 件のコメント
Justin Elstrott
Justin Elstrott 2021 年 1 月 13 日
Hi Ben,
I like this solution, but there's a typo in the third line. I believe it should read as:
oldvariables = T.Properties.VariableNames;
newvariables = {'date','Title','Website'};
[~,LOCB] = ismember(newvariables,oldvariables);
newTable = T(:,LOCB);

サインインしてコメントする。

カテゴリ

Help Center および File ExchangeData Type Identification についてさらに検索

製品

Community Treasure Hunt

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

Start Hunting!

Translated by