Package com.btaz.datautil

Examples of com.btaz.datautil.DataUtilException

74
75
76
77
78
79
80
81
82
83
84
85
            if(! ignoreSet.contains(field)) {
                headerFields.add(field);
            }
        }
        if(idCounter != 2) {
            throw new DataUtilException("Invalid header format. The \"id\" field must be present in both files");
        }
        if(headerFields.size() < h1.length || headerFields.size() < h1.length) {
            throw new DataUtilException("Invalid header format. Repeated header fields with the same name");
        }
        return headerFields;
    }
View Full Code Here
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
     */
    public static void reduce(File inputFile, File outputFile, Reducer reducer, KeyComparator comparator)
            throws DataUtilException {
        // validations
        if(inputFile == null) {
            throw new DataUtilException("The inputFile parameter can not be a null value");
        }
        if(outputFile == null) {
            throw new DataUtilException("The outputFile parameter can not be a null value");
        }
        if(reducer == null) {
            throw new DataUtilException("The reducable parameter can not be a null value");
        }
        if(comparator == null) {
            throw new DataUtilException("The comparator parameter can not be a null value");
        }

        // reduce all data
        FileInputStream inputStream;
        OutputCollector collector = new OutputCollector(outputFile);
        QueueReader queueReader;

        // Aggregate all items matching the comparator and call the Reducable callback
        try {
            inputStream = new FileInputStream(inputFile);
            BufferedReader br = new BufferedReader(new InputStreamReader(inputStream,
                    DataUtilDefaults.charSet));
            queueReader = new QueueReader(br);

            // reduce data
            String prev = null;
            String curr;
            ArrayList<String> items = new ArrayList<String>();
            while(true) {
                curr = queueReader.readLine();
                if(curr == null) {
                    // no more data
                    reduceItems(collector, reducer, items);
                    break;
                } else if(prev == null) {
                    // first row in a new batch
                    items.add(curr);
                    prev = curr;
                } else if(comparator.compare(prev, curr) == 0) {
                    // same keys
                    items.add(curr);
                    prev = curr;
                } else {
                    // different keys
                    queueReader.push(curr);
                    reduceItems(collector, reducer, items);
                    items.clear();
                    prev = null;
                }
            }
            collector.close();
            inputStream.close();
        } catch (IOException e) {
            throw new DataUtilException(e);
        } catch (MapReduceException e) {
            throw new DataUtilException("Irrecoverable reduce operation", e);
        }
    }
View Full Code Here
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
     * @param comparator comparator to use for sorting
     */
    public static List<File> sort(File sortDir, List<File> files, Comparator<String> comparator) {
        // validations
        if(sortDir == null) {
            throw new DataUtilException("The sort directory parameter can't be a null value");
        } else if(!sortDir.exists()) {
            throw new DataUtilException("The sort directory doesn't exist");
        }
        if(files == null) {
            throw new DataUtilException("The files parameter can't be a null value");
        }

        // sort files
        List<File> sortedFiles = new ArrayList<File>();
        for(File file : files) {
            FileInputStream inputStream = null;
            BufferedWriter writer = null;

            try {
                // readLine part into memory
                ArrayList<String> list = new ArrayList<String>();
                inputStream = new FileInputStream(file);
                BufferedReader br = new BufferedReader(new InputStreamReader(inputStream, DataUtilDefaults.charSet));
                String line;
                while((line = br.readLine()) != null) {
                    list.add(line);
                }
                inputStream.close();
                inputStream = null;

                // sort
                Collections.sort(list, comparator);

                // write sorted partial
                File sortedFile = File.createTempFile("sorted-", ".part", sortDir);
                sortedFiles.add(sortedFile);

                writer = new BufferedWriter(new OutputStreamWriter(new FileOutputStream(
                        sortedFile.getAbsoluteFile(), false),DataUtilDefaults.charSet));

                for(String item : list) {
                    writer.write(item + DataUtilDefaults.lineTerminator);
                }
                writer.flush();
                writer.close();
                writer = null;
            } catch (FileNotFoundException e) {
                throw new DataUtilException(e);
            } catch (UnsupportedEncodingException e) {
                throw new DataUtilException(e);
            } catch (IOException e) {
                throw new DataUtilException(e);
            } finally {
                if(inputStream != null) {
                    try {
                        inputStream.close();
                    } catch (IOException e) {
View Full Code Here
14
15
16
17
18
19
20
21
22
23
     * @param dir directory
     * @param extension filename extension
     */
    public static void deleteFilesByExtension(File dir, String extension) {
        if(extension == null) {
            throw new DataUtilException("Filename extension can not be a null value");
        }
        FilenameFilter filter = new FileExtensionFilenameFilter(extension);
        FileDeleter.deleteFiles(dir, filter);
    }
View Full Code Here
27
28
29
30
31
32
33
34
35
36
     * @param dir directory
     * @param regex regular expression
     */
    public static void deleteFilesByRegex(File dir, String regex) {
        if(regex == null) {
            throw new DataUtilException("Filename regex can not be null");
        }
        FilenameFilter filter = new RegexFilenameFilter(regex);
        FileDeleter.deleteFiles(dir, filter);
    }
View Full Code Here
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
     * @param filter filename filter, if null then delete all files
     */
    public static void deleteFiles(File dir, FilenameFilter filter) {
        // validations
        if(dir == null) {
            throw new DataUtilException("The delete directory parameter can not be a null value");
        } else if(!dir.exists() || !dir.isDirectory()) {
            throw new DataUtilException("The delete directory does not exist: " + dir.getAbsolutePath());
        }

        // delete files
        File [] files;
        if(filter == null) {
            files = dir.listFiles();
        } else {
            files = dir.listFiles(filter);
        }
        if (files != null) {
            for(File file : files) {
                if(file.isFile()) {
                    if(!file.delete()) {
                        throw new DataUtilException("Failed to delete file: " + file.getAbsolutePath());
                    }
                }
            }
        }
    }
View Full Code Here
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
     * @return <code>List</code> of <code>FilePart</code> items
     */
    public static List<File> split(File splitDir, File inputFile, long maxBytes, boolean skipHeader) {
        // validations
        if(splitDir == null) {
            throw new DataUtilException("The split directory parameter can not be a null value");
        } else if(!splitDir.exists()) {
            throw new DataUtilException("The split directory is not a valid path");
        }
        if(inputFile == null) {
            throw new DataUtilException("The input file parameter can not be a null value");
        } else if(!inputFile.exists()) {
            throw new DataUtilException("The input file doesn't exist");
        }
        if(maxBytes < DEFAULT_MIN_BYTES) {
            throw new DataUtilException("Invalid max bytes specification: " + maxBytes + ", minimum is: "
                    + DEFAULT_MIN_BYTES);
        }

        ArrayList<File> splitFiles = new ArrayList<File>();

        // open file
        FileInputStream inputStream = null;
        LinkedList<String> rows = new LinkedList<String>();
        try {
            // create FilePart
            inputStream = new FileInputStream(inputFile);
            BufferedReader br = new BufferedReader(new InputStreamReader(inputStream, "UTF-8"));
            long charsTotal = 0;
            String line;

            int maxChunk = 60;
            int rowNumber = 0;
            while((line=br.readLine()) != null) {
                rowNumber += 1;
                if(skipHeader && rowNumber == 1) {
                    continue;
                }
                if(storageCalculation(rows.size(), line.length(), charsTotal) > maxBytes) {
                    if(maxChunk-- <= 0) {
                        break;
                    }
                    // we have to stop or we may exceed max part size
                    File splitFile = writePartToFile(splitDir, rows);
                    splitFiles.add(splitFile);
                    rows.clear();
                    charsTotal = 0;
                }
                rows.add(line);
                charsTotal += line.length() + DataUtilDefaults.lineTerminator.length();
            }
            // no more data, finish up the final part
            if(charsTotal > 0) {
                File splitFile = writePartToFile(splitDir, rows);
                splitFiles.add(splitFile);
            }

            inputStream.close();
            inputStream = null;

        } catch (FileNotFoundException e) {
            throw new DataUtilException(e);
        } catch (UnsupportedEncodingException e) {
            throw new DataUtilException(e);
        } catch (IOException e) {
            throw new DataUtilException(e);
        } finally {
            if(inputStream != null) {
                try {
                    inputStream.close();
                } catch (IOException e) {
View Full Code Here
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
            }
            writer.flush();
            writer.close();
            writer = null;
        } catch (UnsupportedEncodingException e) {
            throw new DataUtilException(e);
        } catch (FileNotFoundException e) {
            throw new DataUtilException(e);
        } catch (IOException e) {
            throw new DataUtilException(e);
        } finally {
            if(writer != null) {
                try {
                    writer.close();
                } catch (IOException e) {
View Full Code Here
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
     */
    public static List<File> map(File workDir, List<File> inputFiles, Mapper mapper)
            throws DataUtilException {
        // validations
        if(workDir == null) {
            throw new DataUtilException("The workDir parameter can not be a null value");
        }
        if(inputFiles == null) {
            throw new DataUtilException("The inputFiles parameter can not be a null value");
        }
        if(mapper == null) {
            throw new DataUtilException("The mappable parameter can not be a null value");
        }

        // setup output collector
        OutputCollector collector = new OutputCollector(workDir, "map");

        // map all data
        try {
            for(File file : inputFiles) {
                FileInputStream inputStream;

                try {
                    inputStream = new FileInputStream(file);
                    BufferedReader br = new BufferedReader(new InputStreamReader(inputStream,
                            DataUtilDefaults.charSet));

                    // map data
                    String row;
                    while((row = br.readLine()) != null) {
                        mapper.map(row, collector);
                    }

                    inputStream.close();
                } catch (IOException e) {
                    throw new DataUtilException(e);
                } catch (MapReduceException e) {
                    throw new DataUtilException("Irrecoverable map operation", e);
                }
            }
        } finally {
            collector.close();
        }
View Full Code Here
66
67
68
69
70
71
72
73
74
75
76
77
78
79
     * @return file new file
     * @throws DataUtilException data util exception
     */
    public File createFile(File dir, String filename) throws DataUtilException {
        if(dir == null) {
            throw new DataUtilException("The directory parameter can't be a null value");
        }
        try {
            File file = new File(dir, filename);
            return createFile(file);
        } catch (Exception e) {
            throw new DataUtilException("Invalid file: " + filename);
        }
    }
View Full Code Here
TOP

Related Classes of com.btaz.datautil.DataUtilException

  • com.btaz.datautil.arrays.StringArrayMapper
  • com.btaz.datautil.files.comparator.FilePathComparator
  • com.btaz.datautil.files.FileCat
  • com.btaz.datautil.files.FileDeleter
  • com.btaz.datautil.files.FileMapper
  • com.btaz.datautil.files.FileMerger
  • com.btaz.datautil.files.FileMerger$BatchFile
  • com.btaz.datautil.files.FileReducer
  • com.btaz.datautil.files.FileSorter
  • com.btaz.datautil.files.FileSplitter
    • Copyright © 2018 www.massapicom. All rights reserved.
    All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.