Function gives ans=0 when comparing text

Bad title i know, but here is the problem: I have made a function that should take a social security number as an input and then return the persons name and income and taxes paid. When i use and input it returns the answer 0. The files are imported from excel and made into cell, so i guess the problem lies within the strcmp. what can i use when i compare numbers?

global Citizens Income Taxes
Citizens=[];
Income=[];
Taxes=[]
%Importing:
[~,~,Citizens]=xlsread('Citizens.xlsx');
[~,~,Income]=xlsread('Income.xlsx');
[~,~,Taxes]=xlsread('Taxes.xlsx');
Function file:

function [Name,personalIncome,taxespaid]=socialsecuritylook(serialnumber)
    global Citizens;
    global Income;
    global Taxes;
    Name=0;
    personaliIncome=0;
    taxespaid=0;
    serialnumber=0;

      for i=2:size(Citizens,1)
          if strcmp(Citizens{i,1},serialnumber);
              Name = Citizens{i,2};
          end
      end
      for i=2:size(Income,1);
          if strcmp(Income{i,1},serialnumber)
              personalIncome = {i,2}
          end
      end
  end

original excel files:

Citizens file:
    Security number     Name       Gender       Region
    103920482         Hans Kale       M           10
    492012903         Sara Green      F           2

 Income file:
  Security number      Income
  103920482            72000
  492012903            54350

Taxes file:
Region   basic tax   tax limit      higher rate tax
  10        25         50000              45
  2         20         40000              50

thanks in advance

[dpb]

Need to see the actual data storage structure and sample to really tell.

[Stephen23]

Some "features" of this code might need to be thought about a bit:

• Using global is not recommended, and makes debugging much more complicated.
• strcmp where one input is a numeric: what is that intended to do?
• The loops do not use indexing in the allocation, so only the last loop iteration values will be used.
• Using numeric for default value of string data.

[Chriss]

I have added more to the text, hope it helps. I use the global command as i should be able to change taxes in another function and add new people to the regions. I know strcmp doesn't work but haven't found another way yet (have also tried if serialnumber==Citizens(:,1)).

[dpb]

The "data storage structure" in question is how you've read it in, specifically. Oh, I see you did include the xlsread call. That with the file, does give enough info to know what Matlab will do with it...

The SSN lookup is then

>> SSN=103920482;    % pick a matching value so will not fail...
>> find(cell2mat(citz(2:end,1))==103920482)
ans =
   1
>>

Even though you read the raw input, xlsread converts the values in the cells to whatever is its internal form--the SSN values are integers so they're a cell array, the content of which (excepting for the header that's of no interest) is double. Thus the string searches are doomed to failure.

I show a "more better" way using the table class below.

Answer 1

[dpb]

Read the data files as tables and use categorical variables; will make things downstream much simpler.

>> inc=readtable('incomes.dat');
>> inc.SSN=categorical(inc.SSN)
inc = 
     SSN       Income
  _________    ______
  103920482    72000 
  492012903    54350 
>> cit=readtable('citizens.dat');
>> cit.Name=categorical(cit.Name);
>> cit.Gender=categorical(cit.Gender)
cit = 
     SSN          Name       Gender    Region
  _________    __________    ______    ______
  103920482    Hans Kale     M         10    
  492012903    Sara Green    F          2    
>> tax=readtable('taxes.dat')
tax = 
  Region    Basic    Limit    High
  ______    _____    _____    ____
  10        25       50000    45  
   2        20       40000    50  
>>

Then you can write something like--

function [name,inc,taxes]=SSNlookup(SSN,CIT,TAX,INC)
  name=CIT.Name(CIT.SSN==SSN);
  inc=INC.Income(INC.SSN==SSN);
  taxes=figuretax(inc,CIT.Region(CIT.SSN==SSN),TAX);
end

function tax=figuretax(inc,region,TAX)
  taxTable=TAX(TAX.Region==region,2:end);
  tax=min(inc,taxTable.Limit)*taxTable.Basic/100;
  if inc>taxTable.Limit
    tax=tax+rem(inc,taxTable.Limit)*taxTable.High/100;
  end
end

which can be used as--

>> ssn=input('SSN, please? ','s');  % input SSN as text
SSN, please? 103920482
>> [n,i,t]=SSNlookup(ssn,cit,tax,inc)
n = 
   Hans Kale 
i =
     72000
t =
     22400
>>

ADDENDUM/ERRATUM

Fixed edit/copy of lookup for income to use INC table; had copied previous line for name and changed it but not CIT-->INC initially.

Also oversight of not using string argument 's' for the input SSN; categorical variable matching works directly on the string so don't need to cast the numerical value as had initially.

END ADDENDUM now back to regularly-scheduled programming

The key things to note are:

1. No globals in sight; can only update the tables when not using them anyway,
2. The lookup is handled by matching in the tables in very readable fashion

[Stephen23]

+1 for a neat and robust solution

[dpb]

Thanks...I've come to really like categorical and the table is really good at simplifying lookup operations in conjunction with it.

I can tell if one had really, really large databases it may have some performance issues, but most cases are probably within its grasp; at least until shown otherwise.

Answer 2

[the cyclist]

MATLAB variable names are case-sensitive. I notice you have a variable called income, and another called Income. (This is super-confusing.) It looks like you set the value of income to zero, never change that, and then output it.

[Chriss]

The Income is my file, income is the persons income i am trying to display. i could also call that one personalIncome

Answer 3

[Chriss]

I found out the mistake. The security number wasn't saved as a number but as a text. The only thing wrong was the if statement where {} should be ().

for i=2:size(Citizens,1)
          if strcmp(Citizens(i,1),serialnumber);
              Name = Citizens{i,2};
          end
      end

[dpb]

But wouldn't

Name=Citizens(Citizens.SSN==serialnumber.SSN);

be much more legible and simpler to code than a loop and an if... :). Not to mention the ease in going forward.

[Chriss]

Sorry sir, i'm a newbie (; i just try to make it as simple as possible as long as i get the wanted result - and there isn't many wasted calculations.

[dpb]

But, if you don't explore/expand your repertoire, how do you expect to get past newbie status? (And, to be pedantic, the loop and if are wasted calculations... :) )

[Chriss]

Can you perhaps tell my why this function gives the answer "Wrong" and not "Right"?

v=2:length(Taxes);
for i=2:length(Citizens)
  if strcmp(Citizens(i,1),serialnumber)
    region = mat2str(Citizens{i,4})
    if strcmp(Taxes(v,1),region)
      disp('Right')
    else
      disp('Wrong')

given my Citizens file has the following format:

'Serialnumer'    'name'    'm/f'    'Region'
  '154305205'    'Sara'     'f'        5
  '195934023'    'Peter'    'm'        2

and my Taxes file has following format:

'Region'    'Basic'    'Limit'        'High'
  '1'       [27.00]     [55000.00]    [43.00]
  '2'       [22.00]    [55000.00]     [43.00]

[dpb]

Yes, but solving it that way is counter-productive.

Convert the data to a consistent format; if you refuse to use the table class which would be ideal for the purpose here, at least convert the numeric values stored as strings to numerics and then collapse the table into a single array without the header incorporated into the array. Then a simple lookup works altho still the categorical for the region in the TAXES table and the other key variables in the other two would be "the cat's meow!" for use as the id variable.

As illustration of what talking of, using the previous data for the TAXES table I've already got still hanging around in memory...

>> Taxes=cell2mat(table2cell(tax))  % convert to the aforementioned array
Taxes =
        10          25       50000          45
         2          20       40000          50
>> rgn=2;         % the wanted region from a previous operation
>> Taxes(Taxes(:,1)==rgn,:)  % retrieve that region's data
ans =
         2          20       40000          50
>>

Again, no loops in sight nor needed, and certainly not nested loops that will be impossible to expand for further functionality as the job progresses.

This still isn't nearly as easy to write code for or maintain as one has to remember which column of the array holds which variable and use 2D array indexing to retrieve the desired quantity. A table (did I mention the table class yet?) lets you refer to each by name or several by a list of names or using the subscript expression in places where that really is the most efficient.

ADDENDUM

BTW, it takes some effort to "get there from here" having read the raw data originally (which, coincidentally, is the root source of much of your trouble in having to deal with the mishmash of various data types for the same thing and the cell array) -- I loaded your above into cell array t; use your current Taxes array there:

>> Taxes=[cellfun(@str2num,t(2:end,1)) cell2mat(t(2:end,2:end))];
Taxes =
         1          27       55000          43
         2          22       55000          43
>>

and voila!! a much easier dataset to deal with. If you'd used the first num return from xlsread from the git-go, that's the result would have had initially it would appear (altho we've not seen the actual spreadsheet but the presumption is pretty solid based on raw content shown).

And, to show it really isn't difficult (in fact, one might say it's even simpler than the above if just read the doc a little and follow a few examples), to convert the above cell array to the table (didn't I mention the table class yet? :) ), starting with the same t here:

>> Taxes=cell2table(t(2:end,:));          % data to table from cell
>> Taxes.Properties.VariableNames=t(1,:); % set the variable names
>> Taxes.Region=categorical(Taxes.Region) % convert region from cellstr
Taxes = 
  Region    Basic    Limit    High
  ______    _____    _____    ____
  1         27       55000    43  
  2         22       55000    43  
>>

With either of the above, all your searching problems, associated loops and if's just melt away into the night. As the original example shows, particularly the latter really documents the solution in code as well.