MATLAB: Trying to extract different numbers out of structured text

columnsdataextractloopMATLABnumbersrowstext;

Hi guys,

I am trying to extract numbers out of a bunch of strings, where every number is different in each line but they have the same structure like this:

phrases = ["Analyst Actions: Stifel Nicolaus Cuts Apache Price Target to $37 From $40, Maintains Buy Rating"; % to $NEWPT From $OLDPT
          "Analyst Actions: Citigroup Initiates Coverage on Expedia With Buy Rating, $130 Price Target"; % $NEWPT Price Target
          "Johnson & Johnson's PT cut by Credit Suisse Group AG to $159.00. outperform rating. (NYSE:JNJ)"; % to $NEWPT
          "Kroger's equal weight rating reiterated at Stephens. $35.00 PT. (NYSE:KR)"; % $NEWPT PT
          "Analyst Actions: Citigroup Initiates Coverage on Booking Holdings With Buy Rating" % this row has no value and should be "None" "None"
          % more similiar lines in the same structure
          ]
      
      % Extract NewPT from each line
      
      % Extract OldPT from each line
      
      % Write "None" where NewPT or OldPT values are null

I am trying to create two columns – NewPT and OldPT and extract the values as commented above and assign "None" whenever values don't exist

I'll be thankful to anybody who can help me with this.

Thank you!

Best Answer

I believe 'Regular Expression' will extract the target part of string. The following is an example.

% Extract target part of string
newpt = regexp(phrases,'((?<=to \$)\d+\.?\d*|\d+\.?\d*(?= (Price Target|PT)))','match','once');
oldpt = regexp(phrases,'(?<=From \$)\d+\.?\d*','match','once');
% Convert to numerical array
newpt = str2double(newpt);
oldpt = str2double(oldpt);

Related Solutions

MATLAB: Grouping into a table

In a situation like this, I usually plan out the table I want to see in advance, and then create variables that I can use to create that structure in a single call to table. In this case, it looks like you want a single table with two columns, Text and Issuer, which you could create in a single call like this, assuming the variables were created correctly.

tbl = table(Text, Issuer);

Here is how I would create those variables:

% Define all the source documents.
documentList = [
    "Analyst Actions: Citigroup Cuts Fiverr International's Price Target to $22 From $31, Reiterates Neutral Rating";
    "Analyst Actions: JPMorgan Downgrades Beyond Meat to Neutral From Overweight, Cuts Price Target to $134 From $138";
    "Analyst Actions: Credit Suisse Raises Beyond Meat's Price Target to $125 From $115; Keeps Neutral Rating";
    "Analyst Actions: Oppenheimer Initiates Coverage on Beyond Meat With Perform Rating";
    "Analyst Actions: Morgan Stanley Trims GoPro's Price Target to $3.50 From $4.50, Maintains Underweight Rating";
    "GoPro downgraded by Zacks Investment Research to strong sell. $3.75 PT. (NASDAQ:GPRO)";
    "Analyst Actions: BofA Merrill Lynch Upgrades Fiverr International to Buy From Neutral";
    ];
% Define all the issuers.
issuerList = [
    "JPMorgan";
    "Citigroup";
    "Credit Suisse";
    "Oppenheimer";
    "Morgan Stanley";
    "Zacks";
    "BofA"
    "Merrill Lynch";
    "Bank of America"
    ];
% Create variables to store the table columns. The 'Text' column
% is simply the existing document list, but we will have to do
% some work to determine the issuer.
Text = documentList;
Issuer = strings(size(documentList));
% Loop through each issuer.
for k = 1:numel(issuerList)
    
    % Check to see which text(s) contain this issuer.
    ind = contains(Text, issuerList(k));
    
    % Assign this issue to the correct rows in the output table.
    Issuer(ind) = issuerList(k);
    
end
% Create the final table.
tbl = table(Text, Issuer);

If you find it confusing to loop over issuers, you can loop over documents instead, like this:

% Loop through each document
for k = 1:numel(documentList)
    
    % Check each issuer.
    for j = 1:numel(issuerList)
        
        % Assign the issuer only if it's present in the document string.
        if contains(documentList(k), issuerList(j))
            Issuer(k) = issuerList(j)
        end
        
    end
    
end

Either way, this leaves you with a 7x2 table with Text in the first column and Issuer in the second.

MATLAB: Converting loop to vector operation

I don't think that you can vectorize this. But an acceleration is possible:

lastrow = max(find(dataset.price(filteredrows)==price_x & ...
              strcmp(dataset.ordertype(filteredrows), 'buy')));
% Replacement:
lastrow = find(dataset.price(filteredrows)==price_x & ...
               strcmp(dataset.ordertype(filteredrows), 'buy')), 1, 'last');

And:

dataset.last_buy_quantity_price_x = zeros(height(dataset), 1);
...
   if ~isempty(lastrow)
      dataset.last_buy_quantity_price_x(row) = dataset.quantity(lastrow);
   end
end

Compare the ordertype with 'buy' once only before the loop:

isBuy = strcmp(dataset.ordertype, 'buy');

Then inside the loop:

lastrow = find(dataset.price(filteredrows)==price_x & isBux(filteredrows), 1, 'last');

The same works with "dataset.price == price_x" also.

What about removing all or(dataset.price ~= price_x, ~strcmp(dataset.ordertype, 'buy') before the loop? This might reduce the data set massively:

price_x  = 1.1;
m        = or(dataset.price ~= price_x, ...
              ~strcmp(dataset.ordertype, 'buy'));
name     = dataset.assetname(m);
time     = dataset.time(m);
price    = dataset.price(m);
quantity = dataset.quantity(m);
last_buy = zeros(numel(name),1);
for row = 1:numel(name)
   filtered = strcmp(name(row), name) & (time(row) >= time);
   lastrow  = find(filtered, 1, 'last');
   if ~isempty(lastrow)
     last_buy(row) = quantity(lastrow);
   end  
end
dataset.last_buy_quantity_price_x    = zeros(height(dataset), 1);
dataset.last_buy_quantity_price_x(m) = last_buy;

This saves teh time for addressing the struct fields also and I think the simpler the code looks like, the easier is it to maintain.

Best Answer

Related Solutions

MATLAB: Grouping into a table

MATLAB: Converting loop to vector operation

Related Question