Nested Indexing in a Single Line command

Is it possible to extract different data in the following structured string using a single line command?

exp={'George: A5 == BB';...

'Anna: C3 == DD';...

'Smith: E2 == FFF';...

'Ken: G8 == HHHH'};

For example, the obvious method to extract HHHH is the following:

mystr1=split(exp(end),': ',2);

mystr2=split(mystr1(:,2),' == ',2);

mystr2(end)

Can you suggest a single line of command to extract HHHH? Please note all names, right-hand and left-hand sides of the equations can be anything with different lenghts or characters. Only : and == characters are unchanged. The code needs be be universal meaning it sould extract anything (index 3 and 1 for Smith or index 2 and 2 for C3 or index end and end for HHHH) only using a single line command.

[Bob Thompson]

Why do you need to split at ':' and '=='? Splitting at just the equals should work fine.

Also, you can probably use regexp to split at both special characters.

mystr = regexp(exp{end},'\W','split');
mystr{end};

I have not tested this for specific syntax, but the concept should be sound.

[S H]

Thank you Bob but I am wondering if it is possible to combine

mystr=regexp(....)

and

mystr{end}

into a single line command? Something similar to regexp(...){ind1,ind2}?

[Bob Thompson]

So, all you're looking to do is display the answer? This can be done by unsurpressing the regexp command (remove the semicolon at the end of the line).

[S H]

Ok, I will try it. Thanks.

Answer 1

[Steven Lord]

Using the answer from Bob Nbob's comment:

mystr = regexp(exp{end},'\W','split'); mystr = mystr{end};

That's one line (albeit not one command.)

The type of indexing you asked about doesn't exist, at least not directly. You could write an explicit call to subsref to get that in one command, but in my opinion it would be much less clear than writing those two commands.Is there a particular reason why you need it in one command?

If you want to do this to use this in an anonymous function inside a script, and you're using release R2016b or later instead consider creating a local function inside your script file.

[S H]

I am trying to write a memory efficient prgram. It is very long. The code will not be efficient if for every calculation, a temporary variable should be created and then removed.

I think using a function helps me to achieve what I need. At least, functions have their own workspace that is removed by Matlab after using functions.

Thanks.

Answer 2

[Stephen23]

>> C = {'George: A5 == BB'; 'Anna: C3 == DD'; 'Smith: E2 == FFF'; 'Ken: G8 == HHHH'};
>> D = regexp(C,'^(\w+):\s*(\w+)\s*==\s*(\w+)','tokens','once');
>> horzcat(D{:}) % or VERTCAT, depending on your needs.
ans =
  [1,1] = George
  [2,1] = A5
  [3,1] = BB
  [1,2] = Anna
  [2,2] = C3
  [3,2] = DD
  [1,3] = Smith
  [2,3] = E2
  [3,3] = FFF
  [1,4] = Ken
  [2,4] = G8
  [3,4] = HHHH

[S H]

Thank you Stephen but I am wondering if it is possible to combine

D=regexp(....)

and

horzcat(...)

into a single line command? Something similar to regexp(...){ind1,ind2}?

[Walter Roberson]

LastCell = @(C) C{end};

then

LastCell( regexp(C,'^(\w+):\s*(\w+)\s*==\s*(\w+)','tokens','once') )

[S H]

I just want to show a data (for example 'HHH' or 'E2') at a time. Not all the data in a row.

[Walter Roberson]

C = {'George: A5 == BB'; 'Anna: C3 == DD'; 'Smith: E2 == FFF'; 'Ken: G8 == HHHH'};
disp(char(regexprep(C, '^.*==\s*', '')))