%separateNewerDataColumns
% Separates the file luca.newer.csv into individual columns in 
% individual files. The files have the same name as the column
% whose data they store. This is necessary because luca.newer.csv
% is HUGE: 700K rows by 50 columns, but this script needs to be
% run only once.
% NOTE: This script calls the shell and runs a short awk program. 

%Written by Luca Cazzanti
%Copyright 2005
%$Id$

location = 'newer_csv_data';
fileName = fullfile(location, 'luca.newer.csv');

columnNames = {'id', ...
  'cluster_center_index', ... 
  'cluster_center_rank', ...
  'cluster_center_size', ...
  'convergence', 'evalue', ...
  'experiment_astral_ac', ...
  'experiment_contact_order', ...
  'experiment_file', ...
  'experiment_index', ...
  'experiment_percent_alpha', ...
  'experiment_percent_beta', ...
  'experiment_sccs', ...
  'experiment_sequence_key', ...
  'experiment_sequence_length', ...
  'experiment_structure_key', ...
  'ln_e', ...
  'n_decoys_in_outfile', ...
  'nss', ...     
  'nsup', ...
  'prediction_ac', ...
  'prediction_contact_order', ...
  'prediction_file', ...
  'prediction_index', ...
  'prediction_percent_alpha', ...
  'prediction_percent_beta', ...
  'prediction_sccs', ...
  'prediction_sequence_key', ... 
  'prediction_sequence_length', ...
  'prediction_structure_key', ...
  'probability', ...
  'prob2', ...
  'psi1', ...
  'psi2', ...
  'score', ...
  'target', ...
  'zscore', ...
  'zscore_normalized', ...
  'zscore_rank', ...
  'correct_fold', ...
  'correct_superfamily', ...
  'self_match', ...
  'class', ...   
  'cv', ...
  'nt', ...
  'ratio', ...
  'aratio', ...
  'bratio', ...
  'testprob', ...
  'testprob2'};

nColumns = length(columnNames);

for idx = 1:nColumns
  disp([columnNames{idx} '...']);
  columnFileName = fullfile(location, [columnNames{idx} '.txt']);
  awkCommand = ['!awk ''{print $' num2str(idx) '}'' ' ...
    fileName ' > ' columnFileName];
  eval(awkCommand);
end