How to separate rows from an array

Hey guys,

I am fairly new to MatLab and I am trying to figure out how to seperate out certain rows from my main array into a new array based on the contents of my first column.

For  Example 
column 1 column 2 column 3  
1             Cash       5.00
2             Credit     10.00 
3             Debt       1.50

So lets say I wanted to separate out cash and credit and add them to a new array so I can ultimately sum the contents of column 3 together.

Something along the lines of "if column 1 is => 0 & =<2, move whole rows to " array 2"

Thanks for your help.

[Fangjun Jiang]

You can use {}Code format next time.

Answer 1

[Fangjun Jiang]

If you matrix is numerical, it is actually easy.

a=[1            5.00
    2           10.00 
    3           1.50
    1            5.00
    2           10.00 
    3           1.50];

index=a(:,1)>0 & a(:,1)<3;
b=a(index,:)

[Alex]

Wow this is awesome, Thank you. and thank you Honglei for your input too.

[Alex]

OK so I am having another problem.

I am sorting the 1st column correctly but the data from the rest of the respective rows are not copying over with it.

i.e. after using

index=a(:,1)>0 & a(:,1)<3;

b=a(index,:)

it only gives me

1

2

but I need

1 cash 5.00

2 credit 10.00

How do I do this?

[Fangjun Jiang]

What is your a? Can you give it exactly using MATALB syntax?

[Alex]

Ok so I have a structure array with 4 arrays

lets say struct.1, struct.2,struct.3,struct.4

.1 has a list of numbers as double

.2 has a list of 'text'

.3 has a list of number values as cells

.4 has a list of number values as cells

So far I have

index=struct.1(:,1)>0 & struct.1(:,1)<3;

b=struct.1(index,:)

which is working but I need the corresponding values in the other 3 structs to move with the first struct.

I hope this is not too cumbersome

[Fangjun Jiang]

The problem is, you can't have '1' or '2' as the field name for a structure. That's why I ask you to provide the exact a data using valid MATLAB syntax. Any way, here is how it's done if a is a struct array.

%%

a=struct('f1',{1,2,3},'f2',{'Cash','Credit','Debt'},'f3',{5,10,1.5})

Num=[a.f1];

index=Num>0 & Num<3;

b=a(index);

b(1)

b(2)

ans =

f1: 1

f2: 'Cash'

f3: 5

ans =

f1: 2

f2: 'Credit'

f3: 10

[Alex]

ok sorry it's

struct.GL

struct.descrp

struct.preQE

struct.curQE

but my data set is over 250 rows long. is there any short hand for the first step where I have to manually enter the data in the { }?

[Fangjun Jiang]

Then where did you get the data? How do you know if it is struct.GL etc. if you don't even have it in MATLAB yet?

[Alex]

I have a loaded spread sheet named bs_data (250x4) that I am drawing the data from

[Fangjun Jiang]

use [NumData,TxtData,a]=xlsread('bs_data.xls'), you'll get a as a cell array and then use the following.

%%

a={1,'Cash',5.00

2,'Credit',10.00

3,'Debt',1.50}

Num=[a{:,1}];

index=Num>0 & Num<3;

b=a(index,:)

[Alex]

Ok, nice thanks. But in actuality what would I enter into a={} I have about 30 different line items.

for example..

"Net Income" is = to struct.GL numbers (1< & <15), (which change periodically so I don't want to have to change it manually every time I upload a new file).

[Fangjun Jiang]

What are you talking about? Run the xlsread() and your data will be put in the 'a' variable. The code above is just an example showing how it works on a cell array.

[Alex]

Sorry, I just confused myself. This works perfectly. Thank you so much for your help.

Answer 2

[Honglei Chen]

Here is a simple example:

>> x = rand(3,3)

x =

      0.9055    0.6581    0.2495
      0.2045    0.1836    0.3782
      0.9654    0.3151    0.5356

>> y = x(x(:,1)>0.5,2)

y =

      0.6581
      0.3151

HTH

Answer 3

[Alex]

Here's an example using the text 'credit' and 'cash' to find and sort out what you want easily.

col_2 = {'credit' 'credit' 'cash' 'cash'};
  col_3 = [ 1 2 3 4  ];

      %find the credit rows
      credit_rows = strcmp(col_2,'credit');
      %find indexes of credit row
      credit_index = find(credit_rows == 1);

      %now we have the index of each credit value, so we can sum these, or move them
      only_credit_col = col_3(credit_index);
      total_credit = sum(only_credit_col);

      %remove the credit cells from the original column - this will leave you a column with only cash values
      col_3(credit_index) = [];

      cash_col = col_3;
      cash_tot = sum(cash_col);

results: credit_col = [1 2] total_credit = 3

cash_col = [3 4] total_cash = 7 total_cash = 7

[Fangjun Jiang]

credit_rows is a logic array. It can be used for index. You don't need to generate the credit_index using find()

col_2 = {'credit' 'credit' 'cash' 'cash'};

col_3 = [ 1 2 3 4];

%find the credit rows

credit_rows = strcmp(col_2,'credit');

%now we have the index of each credit value, so we can sum these, or move them

only_credit_col = col_3(credit_rows);

total_credit = sum(only_credit_col);

%remove the credit cells from the original column - this will leave you a column with only cash values

col_3(credit_rows) = [];

cash_col = col_3;

cash_tot = sum(cash_col);