/* * [CSVISBN.java] * * Summary: convert ISBN10 or ISBN13 in selected columns. * * Copyright: (c) 2012-2017 Roedy Green, Canadian Mind Products, http://mindprod.com * * Licence: This software may be copied and used freely for any purpose but military. * http://mindprod.com/contact/nonmil.html * * Requires: JDK 1.8+ * * Created with: JetBrains IntelliJ IDEA IDE http://www.jetbrains.com/idea/ * * Version History: * 1.0 2012-12-15 initial version */ package com.mindprod.csv; import com.mindprod.common18.EIO; import com.mindprod.hunkio.HunkIO; import com.mindprod.isbn.ISBNValidate; import java.io.EOFException; import java.io.File; import java.io.IOException; import java.io.PrintWriter; import java.nio.charset.Charset; import static java.lang.System.*; /** * convert ISBN10 or ISBN13 in selected columns. *

* Use: java.exe com.mindprod.CSVSIBN somefile.csv 0 3 * * @author Roedy Green, Canadian Mind Products * @version 1.0 2012-12-15 initial version * @since 2012-12-15 */ public final class CSVISBN { /** * how to use the command line */ private static final String USAGE = "\nCSVISBN needs a single filename.csv on the command line followed by " + "0-based numeric columns to convert ISBN10<->ISBN13."; /** * convert awkward characters in selected columns in a CSV file to entities, * constructor. Just create the CSVISBN object.. There are no methods to call. * * @param fileBeingProcessed CSV file to have awkward chars converted to HTML entities. * @param separatorChar field separator character, usually ',' in North America, * ';' in Europe and sometimes '\t' for * tab. * @param quoteChar char to use to enclose fields containing a separator, usually '\"'. Use (char)0 if * you don't want a quote character. * @param commentChar char to use to introduce comments. Use (char) 0 if none. Only one character allowed. * @param encoding encoding of the input and output file. * @param colsToConvert list of columns wanted to have ISBNs converted. * * @throws java.io.IOException if problems reading/writing file */ @SuppressWarnings( { "WeakerAccess" } ) public CSVISBN( final File fileBeingProcessed, final char separatorChar, final char quoteChar, final char commentChar, final Charset encoding, final int... colsToConvert ) throws IOException { final String commentChars = ( commentChar == 0 ) ? "" : String.valueOf( commentChar ); // reader, separatorChar, quoteChar, commentChars, hideComments, trimQuoted, // trimUnquoted allowMultipleLineFields final CSVReader r = new CSVReader( EIO.getBufferedReader( fileBeingProcessed, 64 * 1024, encoding ), separatorChar, quoteChar, commentChars, true, true /* trimQuoted */, true /* trimUnquoted */, true ); final File tempFile = HunkIO.createTempFile( "temp_", ".tmp", fileBeingProcessed ); // writer, quoteLevel, separatorChar, quoteChar, commentChar, trim final PrintWriter pw = EIO.getPrintWriter( tempFile, 32 * 1024, encoding ); final CSVWriter w = new CSVWriter( pw, 0 /* minimal */, separatorChar, quoteChar, commentChar, true ); try { while ( true ) { // process one line each time through the loop. final String[] fields = r.getAllFieldsInLine(); final int fieldCount = r.wasComment() ? fields.length - 1 : fields.length; // convert selected column to entities for selected columns. for ( int source : colsToConvert ) { if ( source < fieldCount ) { final String rawISBN = fields[ source ]; final String isbnWithoutDashes; if ( rawISBN.contains( "-" ) ) { isbnWithoutDashes = ISBNValidate.tidyISBN10or13RemovingDashes( rawISBN ); } else { isbnWithoutDashes = rawISBN; } switch ( isbnWithoutDashes.length() ) { // convert 10 to 13 without dashes case 10: fields[ source ] = ISBNValidate.isbn10To13( isbnWithoutDashes ); break; // convert 13 to 10 without dashes case 13: fields[ source ] = ISBNValidate.isbn13To10( isbnWithoutDashes ); break; default: err.println( " line " + r.lineCount() + " contains invalid ISBN " + rawISBN + " Left as is." ); } } } w.nl( fields, r.wasComment() ); } // end while } catch ( EOFException e ) { out.println( r.lineCount() + " lines converted." ); r.close(); w.close(); HunkIO.deleteAndRename( tempFile, fileBeingProcessed ); } } /** * Simple command line interface to CSVISBN. Corverts selected columns in one csv file whose name is on the * command line. Must have * extension .csv
Use java com.mindprod.CSVISBN somefile.csv 0 1 2 3 ... * Output replaces input. If you want the input, make a copy first. * * @param args name of csv file to convert, followed by zero-based cols to process in desired order. */ public static void main( final String[] args ) { if ( args.length < 2 ) { throw new IllegalArgumentException( USAGE ); } final String filename = args[ 0 ]; if ( !filename.endsWith( ".csv" ) ) { throw new IllegalArgumentException( "Bad Extension\n" + USAGE ); } final File file = new File( filename ); final int[] colsToStrip = new int[ args.length - 1 ]; try { for ( int i = 1; i < args.length; i++ ) { colsToStrip[ i - 1 ] = Integer.parseInt( args[ i ] ); } } catch ( NumberFormatException e ) { throw new IllegalArgumentException( USAGE ); } try { // file, separatorChar, quoteChar, commentChar new CSVISBN( file, ',', '\"', '#', CSV.UTF8, colsToStrip ); } catch ( IOException e ) { err.println(); e.printStackTrace( err ); err.println( "CSVISBN failed " + EIO.getCanOrAbsPath( file ) ); err.println(); } } // end main }