Java代码未处理巨大的CSV文件

Question

I'm new to Java, and I have to write a code that processes a huge CSV files, more specifically to pivot the file and output it in a different way.

When I test the code with couple hundreds of rows it works perfectly! but when I make the file around 700k rows, which it's the case, it outputs a file without any values!

Here is the file schema I'm using:

row_number,Time_float,V1_float,V2_float,V3_float,V4_float,V5_float,V6_float,V7_float,V8_float,V9_float,V10_float,V11_float,V12_float,V13_float,V14_float,V15_float,V16_float,V17_float,V18_float,V19_float,V20_float,V21_float,V22_float,V23_float,V24_float,V25_float,V26_float,V27_float,V28_float,Amount_float,Class_float
0,-1.996583023457193,-0.6942423209592997,-0.04407492457044802,1.6727734992241514,0.973365514375461,-0.24511658354252203,0.34706794516190337,0.19367893831762997,0.0826372794108694,0.33112778320993075,0.08338554524255039,-0.540407035731003,-0.6182957177945313,-0.9960989219768981,-0.32461018632700356,1.6040138389062168,-0.5368328685192513,0.24486345402090126,0.030769932602201018,0.4962820266510568,0.3261180160164485,-0.024923364961491876,0.38285443833968436,-0.17691133433749112,0.1105069205607409,0.2465854429694212,-0.3921704315485548,0.3308916226487169,-0.06378115069750527,0.24496426337017338,-0.04159897836869265
1,-1.996583023457193,0.608496327692869,0.16117591988327584,0.10979710210602639,0.31652292674812704,0.043483352047263765,-0.06181996595114916,-0.06370020974401087,0.07125348305960227,-0.23249418894056792,-0.15334962861450452,1.5800028495181635,1.0660885709726662,0.49141820388511015,-0.14998248182742244,0.6943604186631746,0.5294337538197206,-0.13516996909847068,-0.21876258231230966,-0.17908604930753447,-0.08961086263099204,-0.3073768045456866,-0.880076754448717,0.16220118362512442,-0.5611305498033158,0.3206939009070385,0.2610694754212664,-0.02225567818703569,0.04460751768012477,-0.3424745411051305,-0.04159897836869265
2,-1.9965619655334634,-0.6935004627168448,-0.8115778263099092,1.1694684928233277,0.2682312938426663,-0.3645717858377013,1.3514535859514658,0.6397756379029846,0.20737272949225669,-1.3786753514283783,0.19069961380161685,0.6118297100478469,0.06613661868011089,0.720699852326139,-0.173113888845362,2.5629061849805295,-3.298235372241396,1.3068678794417397,-0.14478999149835625,-2.7785608505208725,0.6809749715218957,0.33763169617058825,1.0633582711230263,1.456319745720107,-1.1380921384645095,-0.6285367205194723,-0.2884467520153837,-0.1371368556920797,-0.18102082710565728,1.1606859252297228,-0.04159897836869265
3,-1.9965619655334634,-0.4933248981469483,-0.11216942463929835,1.1825164508596404,-0.6097266412236745,-0.007468880343610824,0.9361498321757143,0.19207063819763437,0.3160175995113439,-1.2625031722059106,-0.050467953146436886,-0.22189161429591525,0.17837098770941834,0.5101687012441243,-0.30036049360793543,-0.6898374093803747,-1.2092959931757417,-0.8054446421603216,2.3453045221536306,-1.5142049233314772,-0.26985522543866364,-0.1474432966746349,0.007266907401368897,-0.30477654737727394,-1.941027139612268,1.2419037126390184,-0.46021734156146177,0.15539620725948264,0.1861885865396321,0.14053425198451397,-0.04159897836869265
4,-1.9965409076097336,-0.5913297637052182,0.5315410497431424,1.0214116755556233,0.28465540421364943,-0.29501543610441505,0.07199858317395955,0.4793022833609244,-0.22651023121331143,0.7443262870821311,0.6916250322008002,-0.8061465859382944,0.5386266478753359,1.3522443515849518,-1.1680335140679103,0.1913234721415302,-0.5152051215648076,-0.2790807862169683,-0.04556900286007717,0.987037297022314,0.5299387935696488,-0.012839217648641888,1.1000112712256314,-0.2201233960065819,0.2332500879381676,-0.39520164174981287,1.041611299621563,0.5436197963831321,0.6518159160992683,-0.07340334025310606,-0.04159897836869265
5,-1.9965409076097336,-0.21747462336443635,0.5816748868716723,0.7525853776136525,-0.11883332826086528,0.30500895924490323,-0.02231347965260122,0.3849359380956985,0.21795466774987732,-0.5176185728033812,-0.34110109646946285,1.3140464540936228,0.360182114067614,-0.3597915734090633,-0.1430571341620137,0.5655071130706558,0.45845966002880245,-0.06844505898436722,0.08190792006276612,-0.04077665181354113,0.1102154145899529,-0.2835222214491732,-0.7714270019345486,-0.04227284666287563,-0.6132733979473761,-0.4465835771366699,0.21963714625507125,0.6289004813773338,0.24563620173852233,-0.33855641663507124,-0.04159897836869265
6,-1.996498791762274,0.6277951843506749,0.08538910070753938,0.029922970138274736,0.8493831061038468,0.139019583041878,0.20469452783029024,-0.004170268076019058,0.0679975594125977,0.4232179413095267,-0.0911553627796218,-1.3881568003685008,-0.15394904162357412,-0.7546302429818224,0.174601525806113,0.05478291827131152,-0.5062323708749048,0.0033208458038780472,-0.7301429071537743,-0.05598631103904616,-0.28489532870127926,-0.22833366503144684,-0.373032409163239,-0.24677959261431537,-1.2879725270568725,1.4390366584060272,-0.5334361005147672,0.08549234948827854,0.015655983019391424,-0.33327894285907317,-0.04159897836869265
7,-1.9964356179910847,-0.3289283471917919,0.8586923440261,0.708576234767151,-0.34763105227477603,0.6875116870747386,0.32134542115371234,0.9058597727667212,-3.1882293041144427,0.5601291069522123,1.1474295272213637,-0.6068981847284727,0.2917078257188488,1.7663144960411647,-1.381049034972912,0.7496140133583983,-0.08687803964000758,-1.4389216521167638,-0.42738292340326745,0.39863532675456365,-0.20331695379069067,2.6458886485220905,-1.3992756084115,0.0920853274818958,-1.0727537219105157,-0.7966329656594401,-0.1070748551085608,-2.9901537037868864,-3.2880827705218882,-0.19010747625415497,-0.04159897836869265
8,-1.9964356179910847,-0.45657301681652984,0.17329147453401414,-0.07465262179597631,-0.19177387759003722,1.9341493665366207,2.793594054309211,0.29920595987227944,0.7125916765850463,-0.35685128708200337,-0.3769401127987114,-0.690809076979848,-0.11054073462883784,-0.2876133294193531,0.07756710959787069,-0.3592023119205175,-0.23974544333000244,-0.5884222737443606,0.14169463193440693,0.7006152251433341,0.06840582416474023,-0.09996300675392492,-0.36942472981431596,-0.327055266660248,1.6702690915973604,0.7159429034669824,-0.7966330619330197,0.02910414108912445,0.43142017601489174,0.019392240308193916,-0.04159897836869265
9,-1.9963935021436252,-0.17269774387883696,0.6780049203798014,0.6887814922128964,-0.15692675912027115,0.36179160642264196,-0.1852187109108031,0.5267057814810946,0.05822291589340413,-0.6705871938144513,-0.33691175301635523,0.996966068530707,0.837059523154456,1.0116259929700517,-0.46268060306897046,0.1641174886473323,0.8438820412040243,-0.6369448535966505,0.568708693313514,0.5549770030986336,0.2642433515365529,-0.336155613099156,-0.8732980464093,-0.19343791270246957,-0.6357673003859946,-0.1337734549071511,0.195341606203293,0.6100096992904638,0.25168133078356686,-0.33851643577313184,-0.04159897836869265
10,-1.9963724442198953,0.7398016180995844,-0.7123689085952064,0.6027095930866821,-0.9716078839795883,-1.4282856664438242,-0.4722411589208851,-1.150469214075524,0.04057090066824084,-1.5659577306254295,1.493926862783695,1.1753019928647082,-0.6719776055792052,-0.5163883874999073,-0.09915047715944188,0.2522962688421534,0.03648207425930615,0.29836818807428966,1.0192908179449345,-0.27193462986307226,-0.5022889927521785,-0.012663864860947654,0.43254001227028344,0.0444226909424958,0.826410614996712,0.4822144153958534,-0.26850047275423083,0.10616079403669354,0.04923997135326903,-0.32204432065410754,-0.04159897836869265

and Here is the code:

import com.opencsv.CSVReader;
import com.opencsv.CSVWriter;
import java.io.FileWriter;
import java.io.FileNotFoundException;
import java.io.FileReader;
import java.io.IOException;
import java.util.Arrays;
import java.util.List;

public class test {

    public static void main(String[] args) {
        try {
            //csv file containing data
            CSVReader reader = new CSVReader(new FileReader("/Users/csv-parser/sentiment.csv"));
            CSVWriter writer = new CSVWriter(new FileWriter("/Users/csv-parser/Output.csv"));
            String [] nextLine;
            String [] Target_Columns = new String[] {"row_number", "column_name", "column_value_string", "column_value_float"};
            List<String> Columns = null;
            long start = System.currentTimeMillis();

            int lineNumber = 0;

            while (true) {
                if ((nextLine = reader.readNext()) == null) break;
                lineNumber++;
                if (lineNumber == 1) {
                    writer.writeNext( Target_Columns );
                    Columns = Arrays.asList( nextLine );
//                    System.out.println( Arrays.toString( nextLine ) );
                    continue;
                }
                System.out.println( Arrays.toString( nextLine ) );
                int indexOfRow = Columns.indexOf( "row_number" );
                String row_value = "{ROW_NUMBER},{COl_NAME},{COL_STR},{COL_FL}";
                int i = 1;
                for(String column: Columns){
                    if (column.equals("row_number")) continue;
                    row_value = row_value.replace( "{ROW_NUMBER}", nextLine[0]);

                    if (column.contains( "_string" )) {
//                        System.out.print( "col_str: " + finalNextLines[i] + " " );
                        row_value = row_value.replace( "{COL_STR}", nextLine[i] );
                        row_value = row_value.replace( "{COL_FL}", " " );
                        row_value = row_value.replace( "{COl_NAME}", column.replace( "_string", "" ) );
                    }
                    if (column.contains( "_float" )) {
//                        System.out.print( "col_fl: " + finalNextLines[i] + " " );
                        row_value = row_value.replace( "{COL_STR}", " " );
                        row_value = row_value.replace( "{COL_FL}",  nextLine[i] );
                        row_value = row_value.replace( "{COl_NAME}", column.replace( "_float", "" ) );
                    }


                    i++;
//                    System.out.println( "ROW: " + row_value + " " );
                    writer.writeNext( row_value.split( "," ) );
                    row_value = "{ROW_NUMBER},{COl_NAME},{COL_STR},{COL_FL}";
                }

//                System.out.println("\n");
            }
            writer.close();
            long finish = System.currentTimeMillis();
            System.out.println( "Time elapsed: " + (finish - start));

        } catch (FileNotFoundException e) {
            e.printStackTrace();
        } catch (IOException e) {
            e.printStackTrace();
        }
    }
}

and also, is there a way to make the code read/processes the file concurrently or using multithreads, I mean make it more efficient?

Update:

Here how the output looks like when I process the original file

Answer 1

In these very large files, occasionally you get bad input. The best time to check for nonsense input is in the read, at a record / row level (for CSV files, row is the natural choice).

So put an "if" statement in the iterator loop, and have it crash the program with a good error statement on whatever you deem to be bad input. Who knows, you just might find 400 blank rows between a large number of fully populated rows. Also, you might find special control characters, etc.

Once you feel your input is correct, you can start debugging into the logic of the pivot. Who knows, maybe you found a bug in the library.

Personally, I think it's a problem with the input data, as you indicate the logic works well when using smaller data sets. Odds are there's something weird in the larger data set that you haven't noticed, due to its size.

Answer 2

It's really hard to say what is wrong without a full stack trace because your code has lots of potential problems with performance, JVM memory usage etc. I think it's simpler to give you one of a correct solution; at leats, it works with 700k text file.

import com.opencsv.CSVReader;
import com.opencsv.CSVWriter;

import java.io.FileReader;
import java.io.FileWriter;
import java.io.IOException;
import java.nio.file.Path;
import java.nio.file.Paths;
import java.util.Collections;
import java.util.LinkedHashMap;
import java.util.Map;
import java.util.Set;

public class ReadJSONExample {

    public static void main(String[] args) throws IOException {
        // Move time calculation out of main logic
        long start = System.currentTimeMillis();

        new CsvDataModifier(Paths.get("e:/data.csv"), Paths.get("e:/res.csv")).apply();

        long finish = System.currentTimeMillis();
        System.out.printf("Time elapsed: %d ms", finish - start);
    }
}

final class CsvDataModifier {

    private final Path src;
    private final Path dest;

    public CsvDataModifier(Path src, Path dest) {
        this.src = src;
        this.dest = dest;
    }

    private static final int ROW_NUMBER = 0;
    private static final int COLUMN_NAME = 1;
    private static final int COLUMN_VALUE_STRING = 2;
    private static final int COLUMN_VALUE_FLOAT = 3;
    private static final String[] COLUMN_NAMES = { "row_number", "column_name", "column_value_string", "column_value_float" };

    public void apply() throws IOException {
        // Use `try with recources` to close streams correctly
        try (CSVReader reader = new CSVReader(new FileReader(src.toFile())); 
             CSVWriter writer = new CSVWriter(new FileWriter(dest.toFile()), ',', CSVWriter.NO_QUOTE_CHARACTER)) {
            // key - ordered list of columns in source file
            Map<String, Marker> columnNameFloatMarker = getSourceColumnNamesWithFloatMarker(reader.readNext());
            int posRowNumber = getRowNumberPosition(columnNameFloatMarker.keySet());

            if (columnNameFloatMarker.isEmpty())
                return;

            writer.writeNext(COLUMN_NAMES);

            // Create buffer only once for all lines; do not use string concatenation or replacing
            String[] buf = new String[COLUMN_NAMES.length];

            reader.forEach(values -> {
                buf[ROW_NUMBER] = values[posRowNumber];

                int col = 0;
                // this is just reference to buf[]; if `null` then no output
                String[] resultLine;

                for (Map.Entry<String, Marker> entry : columnNameFloatMarker.entrySet()) {
                    String columnName = entry.getKey();
                    Marker marker = entry.getValue();

                    if ((resultLine = marker.createResultLine(columnName, values[col], buf)) != null)
                        writer.writeNext(resultLine);

                    col++;
                }
            });
        }
    }

    private static final String FLOAT = "_float";
    private static final String STRING = "_string";

    private enum Marker {
        NONE {
            @Override
            public String[] createResultLine(String columnName, String value, String[] buf) {
                return null;
            }
        },
        STRING {
            @Override
            public String[] createResultLine(String columnName, String value, String[] buf) {
                buf[COLUMN_VALUE_STRING] = value;
                buf[COLUMN_VALUE_FLOAT] = " ";
                buf[COLUMN_NAME] = columnName;
                return buf;
            }
        },
        FLOAT {
            @Override
            public String[] createResultLine(String columnName, String value, String[] buf) {
                buf[COLUMN_VALUE_STRING] = " ";
                buf[COLUMN_VALUE_FLOAT] = value;
                buf[COLUMN_NAME] = columnName;
                return buf;
            }
        };

        public abstract String[] createResultLine(String columnName, String value, String[] buf);
    }

    // Source column preprocessing to avoid string comparision; do it only once
    private static Map<String, Marker> getSourceColumnNamesWithFloatMarker(String... columns) {
        if (columns == null || columns.length == 0)
            return Collections.emptyMap();

        Map<String, Marker> map = new LinkedHashMap<>();

        for (int i = 0; i < columns.length; i++) {
            String columnName = columns[i];
            Marker marker = Marker.NONE;

            if (columnName.endsWith(FLOAT)) {
                columnName = columnName.substring(0, columnName.length() - FLOAT.length());
                marker = Marker.FLOAT;
            } else if (columnName.endsWith(STRING)) {
                columnName = columnName.substring(0, columnName.length() - STRING.length());
                marker = Marker.STRING;
            }

            if (map.put(columnName, marker) != null)
                throw new IllegalArgumentException("Column duplication in the source file");
        }

        return map;
    }

    private static int getRowNumberPosition(Set<String> columnNames) {
        int i = 0;

        for (String columnName : columnNames) {
            if ("row_number".equals(columnName))
                return i;
            i++;
        }

        throw new IllegalArgumentException("Source file does not contain 'row_number' column");
    }
}

---

• Input file: data.csv ~432 Mb
• Output file: res.csv ~868 Mb
• Console outputs: Time elapsed: 21058 ms

Answer 3

Your output file is bigger than input file. It's possible that when the code is run against 700k rows it is producing a huge file which is difficult for editors to open.

What's the size of /Users/csv-parser/Output.csv when you run it for 700k rows?

Did you try printing the content of /Users/csv-parser/Output.csv using code rather than opening it using editor?

Answer 4

Have you tried to put it in another thread/task and run it? Maybe it is blocking the main thread...

Check this out: https://winterbe.com/posts/2015/04/07/java8-concurrency-tutorial-thread-executor-examples/