Find numbers that are not in multiple columns

I have three tables:

A1 =
  5×2 table
    x    Var1   
    __   ____
    1    2524
    2    2342
    3    4353
    4    4353
    5    2348

A2 =
  5×2 table
    x    Var2   
    __   ____
    1    6745
    3    5942
    4    9569
    5    2454
    6    1240

A3 =
  5×2 table
    x    Var3   
    __   _____
    2    4252
    3    9345
    4    9235
    5    1246
    6    3965

For each table I want to find and delete the rows that contains a number in the first row that is not in the first rows of both other tables. So I want to end up with this:

A1 =
  3×2 table
    x    Var1   
    __   ____
    3    4353
    4    4353
    5    2348

  3×2 table
    x    Var2   
    __   ____ 
    3    5942
    4    9569
    5    2454

A3 =
  3×2 table
    x    Var3   
    __   ____
    3    9345
    4    9235
    5    1246

How can I do this? Preferably with no loops.

[Adam Danz]

> For each table I want to find and delete the rows that contains a number in the first row that is not in the first rows of both other tables.

Based on the examples you provided (which are helpful to explain the goal, thanks!), it sounds like you want to delete any rows that contain a number in column 1 that is not in the first row, column 1 of the other tables. Otherwise, why is the first row of A3 deleted?

[Dylan den Hartog]

That is correct. Because the number '2' in table 3 is not in table 2, I want to delete it.

[Siddharth Bhutiya]

I believe this is just a dummy example, but is there a reason you have a table with 1 two column variable instead of creating a table with 2 variables? The description of your problem makes it seem that the two columns of your Var1 are separate thing (they just seem to have the same data type).

[Dylan den Hartog]

You are right. I made I mistake. The tables are 5x2. I changed it in the problem description.

Answer 1

[Adam Danz]

It looks like you are deleting all rows that do not contain the Var1 value in all 3 tables. Since all the tables will then have the same Var1 values you could combine them into 1 table. Use innerjoin :

% Demo tables
A1 = table((1:5)', rand(5,1))

A1 = 5×2 table
    Var1     Var2  
    ____    _______

     1      0.84504
     2      0.19125
     3      0.23803
     4      0.32501
     5      0.41045

A2 = table([1;3;4;5;6], rand(5,1)) 

A2 = 5×2 table
    Var1     Var2  
    ____    _______

     1       0.7915
     3      0.85265
     4      0.30851
     5       0.1263
     6      0.32535

A3 = table((2:6)', rand(5,1)) 

A3 = 5×2 table
    Var1     Var2  
    ____    _______

     2      0.37713
     3      0.26054
     4      0.91348
     5       0.9767
     6      0.97345

% Join table A1 and A2 to create A12
A12 = innerjoin(A1,A2,'keys', 'Var1');
% Join A12 with A3 to create A123
A123 = innerjoin(A12, A3, 'keys','Var1'); 

You can rename the variables using A123.Properties.VariableNames = ____

If you'd rather not join the tables

a12mem = ismember(A1.Var1, A2.Var1);
a13mem = ismember(A1.Var1, A3.Var1);
A1new = A1(a12mem & a13mem,:)

A1new = 3×2 table
    Var1     Var2  
    ____    _______

     3      0.23803
     4      0.32501
     5      0.41045

a21mem = ismember(A2.Var1, A1.Var1);
a23mem = ismember(A2.Var1, A3.Var1);
A2new = A2(a21mem & a23mem,:)

A2new = 3×2 table
    Var1     Var2  
    ____    _______

     3      0.85265
     4      0.30851
     5       0.1263

and repeate for the 3rd table.

Answer 2

[Sulaymon Eshkabilov]

Simle solution is a logical indexing approach, something like ...e.g.:

IND = A1.Var1~=A2.Var2;
A1=array2table([A1.x(IND), A1.Var1(IND)], 'variablenames', {'x', 'Var1'});
... 

[Adam Danz]

This does an ordered-comparison. Notice that the first row of table A3 should be removed but none of the other tables contains a 2 in the first row of column 1.

You'd need ismember if I understand the OP correctly.